AI Product Manager · AI Data Strategist
AI Product Manager with 15+ years shipping data- and AI-powered products — scaling platforms from $0.5M to $4M and driving up to $16M in revenue impact across pharma, healthcare, and tech.
I own the full product lifecycle — from discovery and opportunity sizing to roadmap, delivery, and adoption — partnering with engineering, data science, and design to ship AI capabilities that customers and stakeholders actually use.
I specialize in AI/ML product strategy, data and platform strategy, and 0-to-1 through scale execution. I define the problems worth solving, translate them into crisp product requirements and success metrics, and run a hypothesis-led, experiment-driven approach to de-risk ambiguity.
My focus is on responsible, production-grade AI: pragmatic roadmaps, clear guardrails for safety and compliance, and measurable impact on engagement, efficiency, and revenue.
A Claude-powered version of me, grounded on my real career. Ask about the AI products I've built, my product/data background, or whether I'm a fit for your role — then email the real me. Five questions per hour, resetting automatically, so make them count.
Grounded on my real career, not a guarantee of accuracy. For anything specific, email me directly.
Drag to reposition · click a node to highlight its connections · click empty space or Reset to clear
Real products built and shipped across industries — from publishing and healthcare to pharma and consumer tech.
Project names anonymized for client confidentiality · Outcomes from internal program reporting
AI product mining medical, clinical, and field-engagement data to surface new drug insights for MSL and Medical Affairs teams — using NLP and ML to detect trends and knowledge gaps.
AI-driven PLM platform automating provider onboarding, credentialing, claims processing, and compliance tracking — integrating structured healthcare data across enrollment workflows.
Custom Clinical Trial Management System centralizing site selection, enrollment tracking, milestone monitoring, and document control — integrated with EDC, regulatory, and safety systems.
Platform matching customers with vetted chefs based on cuisine preferences, dietary needs, and budget — with streamlined onboarding, scheduling, and secure payments.
Android app store connecting developers and users through a unified storefront with secure payment processing, catalog management, and purchase-based app entitlement at scale.
Predictive analytics product that scores a book's market feasibility before publishing — automating go/no-go decisions using historical sales, genre, and market-trend data.
Personal AI and data products I've built from scratch — exploring what's possible at the frontier of technology.
In the lab — architecture defined, builds in progress. Open to collaborators.
ProblemEnterprises can't trust LLM-only answers about their data — especially in regulated industries that need to prove it to an auditor.
ApproachBuilt a four-agent architecture where deterministic engines (profiling, SQL safety, PII detection) own the ground truth and the LLM only narrates.
ResultAn auditable, HIPAA/GDPR/21 CFR Part 11-mapped data copilot — architecture and roadmap complete, targeting Q3 2026.
Most "AI data copilot" demos ask an LLM to do everything — including things LLMs are bad at, like guaranteeing a SQL query is safe or reliably counting nulls. Enterprises don't just need a plausible-sounding answer about their data; they need to trust it, and in regulated industries, prove that trust to an auditor.
An agentic platform built around one rule: deterministic engines own the ground truth, the LLM only explains, classifies, and narrates. Four agents each pair a real profiling, validation, or detection engine with an LLM interpretation layer on top — a Data Quality Agent (ydata-profiling / Great Expectations) for nulls, duplicates, and schema drift; a Catalog Agent (pgvector + metadata graph) for lineage and semantic search; an Analytics Agent (sqlglot) that only ever executes safe, read-only, LIMIT-bound SQL generated from natural language; and a Governance Agent (Microsoft Presidio) that scans for PII/PHI and maps findings to HIPAA, GDPR, and 21 CFR Part 11. Langflow handles the reasoning for each agent, n8n orchestrates scheduling, routing, and the audit trail, and MCP is the bridge connecting them — no brittle glue-code HTTP calls. The whole pipeline runs against a synthetic pharma/clinical-trial dataset deliberately seeded with real data-quality issues and PII/PHI fields, so it's inspectable end to end rather than a polished demo.
Coming soon, targeting Q3 2026 — architecture and roadmap defined, with the phased build plan documented in the repo.
Problem34,023 trials are registered TERMINATED on ClinicalTrials.gov, and 11.5% of a sampled 200 give zero substantive stop reason — the rest give only a one-line label that takes manual cross-referencing to actually explain.
ApproachBuilt a stateless, single-input web app — paste an NCT ID, get back ranked failure hypotheses pulled only from ClinicalTrials.gov and PubMed, with every claim schema-enforced as Fact, Inference, or Hypothesis.
ResultA live, frictionless tool — no login, no accounts, client-side validation before any network call — targeting sub-20-second results and ≥70% top-hypothesis accuracy against a documented benchmark set.
Clinical trials fail for reasons that are rarely stated plainly. Live ClinicalTrials.gov data puts 34,023 trials at TERMINATED status (5.7% of all 595,847 registered studies) — and combined with WITHDRAWN and SUSPENDED, 52,362 trials (8.8%) never reached completion as planned. A random sample of 200 TERMINATED trials shows 8.0% with no stop reason recorded at all and another 3.5% citing only a generic "business decision" with no underlying cause — 11.5% giving literally zero signal. The remaining 88.5% do name a cause, but most are single-line labels like "low enrollment" or "Covid-19" that don't explain the deeper why without cross-referencing enrollment data, results, and publication history by hand. No existing tool does causal reasoning over this public data: ClinicalTrials.gov and the AACT database expose raw fields with no reasoning, PubMed requires manual cross-referencing, and enterprise tools like Citeline/Trialtrove are comprehensive but priced for institutions and still largely descriptive rather than causal.
ClinicalAI-TrialAnalysis is a single text input: paste an NCT ID and the backend fetches the ClinicalTrials.gov v2 record, resolves linked PubMed publications via NCBI E-utilities, and runs a structured reasoning pass against an explicit 7-category failure taxonomy — recruitment, efficacy, safety/toxicity, funding/business, operational/protocol, strategic/competitive, futility — with few-shot examples and a hidden chain-of-thought scratchpad reasoning through the evidence before committing to a final ranked answer. Every claim across the four output blocks (Bottom Line, Ranked Hypotheses, Evidence, Guardrail) is tagged Fact, Inference, or Hypothesis, enforced by schema validation rather than prompt instruction alone, so an unlabeled claim can never render. The UX is deliberately frictionless: no accounts, no login, no saved history — the product is fully stateless. A malformed NCT ID is rejected client-side before any network call is even made, a well-formed but nonexistent ID gets an explicit "trial not found" state, and if a trial is COMPLETED or still RECRUITING, the tool states plainly that there's no failure to explain instead of fabricating one. If PubMed enrichment fails, it still returns a trial-data-only analysis with a note rather than failing the whole request.
ProblemComparing two companies at an investor-grade level — financials, valuation, moats, risk, outlook — normally means paying for equity research or spending hours across filings and news, and a general chatbot answer skips citations and live data.
ApproachBuilt a bring-your-own-key tool that runs a live, web-searched deep research pass on any two companies and returns a structured, analyst-style report with sourced citations.
ResultA live, no-signup tool — enter two company names and an OpenAI key, get financials, valuation, moats, SWOT, risks, and outlook back in under a minute.
Comparing two companies at an investor-grade level — financials, valuation multiples, competitive moats, risk factors, forward outlook — normally means either paying for equity research or spending hours cross-referencing filings, news, and analyst notes yourself. Ask a general chatbot instead and you get a plausible-sounding paragraph with no live data, no citations, and no way to check if any of it is current.
Company Comparison takes two company names and runs a live, GPT-4o-mini-powered deep research pass with real-time web search, then returns a structured comparison across financials, valuation, moats, SWOT, risks, and outlook, with citations attached to the claims that need them. It's a bring-your-own-key tool by design: the OpenAI key is entered client-side, used only to call OpenAI directly from the browser, and never stored or sent anywhere else — so there's no signup, no backend holding credentials, and no cost to run beyond what the user's own key spends. A single report can take up to a minute since it's doing live research rather than replaying cached training data.
ProblemSMBs and freelancers sign NDAs and MSAs without in-house counsel — a single review costs $1,500–$3,000 and 90–120 minutes at standard legal rates, and generic AI summarizers don't cite where a term came from or how confident they are.
ApproachBuilt a GPT-4o extraction pipeline that pulls the terms that matter per contract type, attaches a confidence score and verbatim source sentence to each one, and grounds a follow-up chat entirely in the uploaded document — no chunking, no guessing.
ResultA live, shipped tool targeting sub-15-minute review time and ≥88% extraction F1 on NDAs, built end-to-end with Claude Code against a full PRD with its own eval suite and HHH readiness criteria.
Business professionals — founders, ops managers, procurement leads — routinely sign NDAs and MSAs without fully understanding what they're agreeing to. Without in-house legal teams, reviewing a single contract takes 90–120 minutes and typically costs $1,500–$3,000 at standard $250–$500/hr legal billing rates, and 43% of SMBs report a commercial dispute that traced back to a misunderstood contract term. Existing options are a poor fit at both ends: enterprise contract-lifecycle tools like DocuSign CLM or Ironclad are built for $50k–$500k contracts, and general tools like ChatGPT produce a plausible-sounding summary with no structured extraction, no page attribution, and no way to tell which parts to trust. Rule-based parsers don't hold up either — a standard confidentiality clause alone shows up in 50+ structurally different phrasings across real contracts, which pushes regex-based extraction past a 30% miss rate.
ContractIQ extracts the 10 terms that matter most in an NDA and the 12 that matter most in an MSA — parties, confidentiality obligations, liability caps, termination clauses, and so on — using GPT-4o in JSON mode at temperature 0.1, with a few-shot prompt built from hand-labeled NDA and MSA examples. Every term comes back with three things a plain summary won't give you: a page number, a 0–100% confidence score that's color-coded and never hidden even below 50% — just flagged with a non-dismissible warning — and the verbatim source sentence it was pulled from, with an expandable "Why?" that answers the trust question directly instead of asking for blind faith. A grounded chat layer answers plain-English follow-up questions using the full contract text as context on every turn, with a system prompt that treats "I cannot find this in the document" as a correct answer rather than a failure, and a mandatory page citation on every response. Hallucination is treated as the top trust risk end to end: deterministic decoding settings, a JSON-parse retry loop, an automated regression test confirming the model still says "not found" on off-document questions, and a monthly calibration check against a legal-SME-labeled set plus the CUAD dataset (13,000+ annotations across 510 contracts) — with a UI warning if miscalibration crosses 15%. The whole build is scoped against explicit production-readiness bars before calling anything launched: ≥88% F1 on NDA extraction, ≥85% on MSA, sub-30-second P95 latency, and a correction rate under 12% in any rolling 7-day window. It was developed end-to-end using Claude Code as the primary development partner — a direct test of how far an agentic coding workflow can carry a real product, from a 28-page PRD with its own HHH (Helpful/Honest/Harmless) readiness framework all the way to a live, shipped app.
ProblemStandard RAG tools — NotebookLM, ChatGPT uploads, most "chat with your docs" pipelines — retrieve chunks and generate an answer, but the model rediscovers knowledge from scratch on every question. Nothing accumulates.
ApproachBuild an agent that incrementally writes and maintains a persistent, interlinked markdown wiki between me and my raw sources.
ResultA personal knowledge base that compounds instead of resets, with the LLM absorbing the maintenance burden that kills most human-run wikis — architecture defined, build targeting Q4 2026.
Most "chat with your documents" tools share the same flaw: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer — but it's rediscovering knowledge from scratch on every question. Read fifty articles on a topic and the fifty-first question still starts from zero. Vannevar Bush described the fix back in 1945 with the Memex — a personal, curated knowledge store with associative trails between documents — but the part he couldn't solve was who does the maintenance. Humans abandon wikis because the upkeep grows faster than the value they return.
Second Brain flips the RAG model around. Instead of chunk-and-retrieve, the LLM incrementally builds and maintains a structured wiki that sits between the user and the raw sources — summarizing new material, updating entity and concept pages, flagging contradictions with earlier claims, and strengthening the synthesis over time. The architecture runs on three layers: immutable raw sources the LLM reads but never edits, an LLM-owned wiki of markdown pages, and a schema doc that defines structure and conventions and co-evolves with the wiki. Three operations drive it — Ingest (a single source can touch 10–15 wiki pages: summary, index, entity updates, log entry), Query (search the index, read the relevant pages, synthesize a cited answer, and file good answers back as new pages so explorations compound too), and Lint (a periodic health check for contradictions, stale claims, and orphaned pages). At moderate scale — roughly a hundred sources, a few hundred pages — a plain index.md file replaces the need for embedding-based retrieval entirely. The workflow: Obsidian as the IDE, the LLM as the programmer, the wiki as the codebase, with the whole thing living in a git repo for free version history and branching.
Coming soon, targeting Q4 2026 — concept and architecture defined, with the wiki schema and Ingest/Query/Lint workflow being tailored for personal knowledge management.
ProblemMid-market property managers are stuck choosing between error-prone spreadsheets and $25K–$500K+/yr enterprise lease platforms.
ApproachShipped the core value first as a Claude Skill — a lease compliance assistant usable instantly, with no deployment project needed.
ResultZero marginal cost and minutes to start, targeting the underserved "moveable middle" — backed by real market research before any code.
Commercial property managers and multi-site occupiers currently have two bad options: spreadsheets, which are fast to start but error-prone and blind to upcoming deadlines until it's too late, or enterprise/mid-market lease platforms (Yardi, MRI, Visual Lease, LeaseQuery), which are accounting-first, built for ASC 842/IFRS 16 journal entries, expensive ($25K–$500K+/yr), and slow to implement (weeks to 18 months). Nobody has built a lightweight, compliance-and-deadline-first tool for the "moveable middle" — mid-market property managers and franchise/multi-unit occupiers with 20–500 leases who care more about renewal windows, permits, insurance certificates, and jurisdictional obligations than right-of-use asset calculations.
Rather than building a standalone SaaS product from scratch, this ships the core value first as a Claude Skill — a SKILL.md that turns Claude into a lease compliance assistant, runnable immediately inside claude.ai, Claude Code, Claude Cowork, or the API, with no deployment or implementation project. Given lease documents and a portfolio spreadsheet, the skill abstracts key lease terms into a structured format, tracks critical dates and obligations (renewals, notice-to-vacate windows, insurance certificate expirations, permit renewals), surfaces jurisdiction-specific regulatory risk that pure accounting tools miss (zoning, ESG/emissions disclosure laws like NYC Local Law 97, ADA/fire-code items), generates compliance-ready reports, and answers natural-language portfolio questions — while explicitly flagging low-confidence, non-standard clauses for human review rather than silently guessing.
ProblemProduct teams spend a full week turning a business question into an answer — by delivery, the context has already moved on.
ApproachBuilt an 18-agent pipeline that frames hypotheses, runs automated analysis, validates findings across four layers, and builds the deck.
ResultAnalyses that took weeks now take minutes — one analyst effectively becomes three, with 80% less time on execution work.
Product teams spend weeks running analyses that should take days. A business question gets asked Monday. The analyst spends Tuesday-Wednesday exploring data, Thursday writing SQL and building charts, Friday documenting findings, and Monday morning delivering a presentation. By then, the business context has shifted, the question has evolved, and the insight arrives too late to act on.
AI Analyst v2 is an AI-powered research partner that handles the execution layer while you keep the judgment layer. You ask a business question. The system runs an 18-agent pipeline that frames your question into testable hypotheses, analyzes your data with automated exploration, segmentation, cohort analysis, and root-cause investigation, validates findings across four layers (structural, logical, business rules, paradox detection), tells the story with a narrative arc and Storytelling with Data charts, and builds the deck — a branded, presentation-ready slide deck with speaker notes and export-ready assets.
ProblemMarketing teams waste time running fragmented, manual ad campaigns across Google, Meta, LinkedIn, and TikTok.
ApproachBuilt an AI-powered automation platform that generates ad copy, configures targeting/budget, and optimizes campaigns in real time.
Result70% faster campaign setup and 35–45% better ad spend ROI, with $200K–$500K in projected annual savings per enterprise.
Marketing teams waste significant time creating, testing, and optimizing ad campaigns across multiple platforms (Google, Meta, LinkedIn, TikTok). Current workflows are fragmented—requiring manual ad copy creation, targeting configuration, budget allocation, and performance monitoring across disparate tools. This inefficiency leads to missed optimization opportunities, delayed market response, and reduced ROI focus. Teams need a unified, intelligent system to accelerate campaign creation and optimization.
Ad Genie is an AI-powered advertising automation platform that streamlines the entire campaign lifecycle. Using Claude AI and machine learning, it auto-generates compelling ad copy variants, intelligently configures targeting and budgets, monitors performance in real-time, and continuously optimizes campaigns for maximum ROI. The platform integrates seamlessly with Google Ads, Meta, LinkedIn, and TikTok, creating a unified advertising control center that reduces manual work and accelerates campaign velocity.
ProblemTeams lose alignment across scattered PM tools with no AI-powered visibility into priorities or dependencies.
ApproachBuilt an AI-powered Kanban platform with a backlog-aware assistant, real-time sync, and secure multi-tenant workspaces.
ResultA live, shipped SaaS product — teams get instant natural-language answers about their own backlog instead of digging through boards.
Modern teams struggle with scattered project management tools that lack intelligent collaboration features. Teams lose alignment across multiple platforms, lack AI-powered insights into priorities and dependencies, spend time context-switching between tools, and can't quickly query their backlog or create issues from natural language. Project tracking becomes an overhead burden instead of a tool that accelerates shipping.
ProjectPilot AI is an AI-powered project management platform combining visual Kanban boards, real-time collaboration, and an intelligent AI assistant. Features include drag-and-drop Kanban boards with unlimited columns, AI assistant that understands your entire backlog and creates issues from natural language, real-time sync powered by Supabase Realtime, secure multi-tenant team workspaces with granular role management (owner, admin, member), and simple flat-rate pricing per workspace (not per seat). The AI knows every issue, priority, and assignee — enabling instant answers to "What are the urgent issues?" and "Who has the most work assigned?"
Where I've built products and driven impact.
Continuous learning across product, technology, and strategy.
Amazon Web Services · Q3 2026
MIT Professional Education · Q1 2024
MIT Management Executive Education · Q1 2024
Scrum Alliance · Q1 2024
SAFe by Scaled Agile · Q1 2024
Anthropic Education · Q3 2026
Maven · Q3 2026
Maven · Q3 2026
Maven · Q3 2026
Maven · Q3 2026
Scrum Alliance · 2026
Amazon Web Services · Q1 2024
Pendo · Q1 2024
Scrum Alliance · Q1 2024
Scrum Alliance · Q1 2024
The tools I use to research, design, build, and ship products.
Open-source Claude AI skills and sub-agents built for product managers. Free to use, fork, and contribute.
Curated Claude AI skills and sub-agents for product managers — market research, PRD writing, evaluation, user research, mockup generation & more. Free to use and contribute.
Q1'26
Full documentation & usage guide for all skills
Q1'26
SkillCompetitive analysis & market intelligence skill
Q2'26
SkillEvaluate and score product requirements docs
Q1'26
SkillGenerate structured product requirements documents
Q2'26
SkillSynthesize user interviews & research insights
LinkedIn recommendations from colleagues and clients.
Chandrashekhar Hajare
Validation Lead
Prag Ravichandran
Founder | Salesforce MVP | Certified Architect | 10x Dreamforce Speaker
Sri Kumaran Thirupathy
Platform & Product Leader | Data Thinker | Curious Mind in Tech & Beyond
Carolyn Hudson, MBA
Program Director, Healthfirst
Vivek Thomas
Senior Manager, Data Informatics
Muthamizhselvan Kalidoss
Senior Manager, Cloud & Infrastructure Services
Imthias Shaffiullah
Software Engineer, Spotline, Inc.
Rajesh Sundarraj
Group Technical Specialist, HCL Technology
Vigneshwaran Sivaprakasam
Group Project Manager, HCL America Inc.
Sneha Gayathri C
QA Engineer, Life Sciences Domain
Ravi Vishwas
Enterprise Technology & ERP Senior Consultant
Giridhar Golkonda
IT Leader & Enterprise Architect
Peter Dudek
Pega Certified Senior System Architect
Paul Atherton
Manager Provider Data Management, Healthfirst
Denise Hurley-Forgach
Senior Manager, Utilities Industries
Narayanan Ramaswamy
Sr. Client Partner @ Saama | Health Sciences
Thukkaram Periasamy
Associate Manager, Accenture
What I've shipped and updated most recently on my GitHub.
Launched "Ask My Twin" — an AI chatbot grounded on my real career, built with Claude and Cloudflare WorkersFeature
Added a command palette (Cmd+K) for instant navigation across the siteFeature
Shipped the Enterprise AI Data Copilot Suite build card and opened it up for collaboratorsBuild
Published the Compliance Lease Tool with a live GitHub repo and mockupBuild
Added new recommendations from colleagues at HCL and SpotlineUpdate
What I've been sharing recently on LinkedIn.
Argued that global team breakdowns get misdiagnosed as culture problems — the real tell is regional leaders with high trust but low influence, and the fix is redesigning decision rights, not fixing cultureArticle
Shipped ClinicalAI-TrialAnalysis — paste an NCT ID, get ranked failure hypotheses from Claude with a self-argued counter-argument on each one, and a hard schema gate that rejects any response with an unlabeled claimAchievement
Broke down why 45% of CEOs see employee resistance to gen AI — competence, autonomy, and relatedness needs going unmet — and the AWARE framework for closing the trust gapArticle
Wrote up how I built the Company Comparison tool — BYOK with no backend, a web-search-capable model for live citations, and a hand-rolled markdown parser plus D3 relationship graph instead of off-the-shelf librariesAchievement
Called out the blind spot in agentic AI roadmaps — infrastructure — where vector databases and legacy storage built for human-paced traffic buckle under machine-speed query volumeArticle
Shipped ContractIQ end-to-end with Claude Code — walked through the step-by-step, from writing the PRD first to defining an HHH launch bar before calling it production-readyAchievement
Named the "two-organizations problem" — the gap between the reported org (dashboards, board decks) and the lived org (what teams actually experience) — and why it widens fastest around AI rolloutsArticle
Broke down "sales debt" — taking on poor-fit customers to hit short-term revenue — and why the bill comes due in shrinking margins, customization-eaten roadmaps, and churn nobody wantedArticle
Argued that two decades of decomposing firms into modular units left companies great at splitting apart and bad at recombining — and that the real advantage is turning recombination into a managerial "handshake," not just a technical interfaceArticle
Unpacked why a Fortune 100 exec's Asia-based team stayed quiet on strategy — not a culture problem but a systems one, and why fixing decision rights, not running a culture audit, is the real leverArticle
Broke down why a two-person team — one domain expert, one AI engineer — is the new baseline for building products that used to need eight, and why incumbents must re-architect workflows before automating themArticle
Broke down AT&T CEO John Stankey's HBR IdeaCast take that most AI efficiency gains get competed away — the durable edge is marrying proprietary data with AI, not "meat and potatoes" automationArticle
Reacted to an HBR piece on Goldman Sachs' CIO reframing the "10% AI can't do" question — the real shift is operator to supervisor of agents, and AI transformation only works once data and evals are in orderArticle
Announced the "Enterprise AI Data Copilot Suite" — an agentic platform where deterministic engines hold the ground truth and the LLM handles explanation, classification, and narrationUpdate
Shipped "ProjectPilot AI" — a full-stack SaaS project management platform with Kanban boards, AI task intelligence, and a workspace-scoped Claude assistant, from zero to production in daysAchievement
Demoed "Ad-Genie" at the end of a 6-week Agentic AI PM cohort — an ad-creative co-pilot built with a PRD, eval framework, and pricing model from scratchAchievement
Reflected on treating pricing as a product decision, not a finance one — mapped Ad-Genie against a fixed/variable, activity/outcome-based frameworkArticle
Explored the real difference between a co-pilot and an agent — initiative — and why Ad-Genie is deliberately scoped as the former for nowArticle
Learned the HHH framework — Helpful, Honest, Harmless — for evaluating generative AI output where there's no single "right" answerArticle
Reframed AI roadmaps as sequencing risk instead of features — validate the riskiest assumption before earning the right to build the next thingArticle
Kicked off an Agentic AI PM cohort with a 2x2 framework for deciding where AI actually belongs — automate, quality-control, co-pilot, or human-firstNew Start
Published my first public GitHub repo — Claude Skill Builder Collections, a growing library of reusable Claude AI skills for product teamsNew Start
Built "ScienceAI" — a RAG-powered research assistant that lets scientists talk to the literature and turns findings into audio summaries or podcastsNew Start
Completed "Building Agentic AI Applications with a Problem-First Approach" on Maven — 35+ hours covering evals, prompt engineering, and multi-agent pipelinesAchievement
Completed the "Become an Agent-Native Operator in 1 Day" OpenClaw Bootcamp — covering agentic workflows, MCPs, and eval loopsAchievement
Passed the AWS Certified Generative AI Developer – Professional (Early Adopter) certificationAchievement
Reach & Engagement
12.0K
impressions · last 90 days
4.0K members reached
3.1 engagements / day avg
Reach & Engagement
Follower Growth
▲ +2,195
new followers · last 90 days
7K+ total followers
👋 Psst, flip me for moreFollower Growth
Post Performance
2,085
most impressions · single post
1.0% engagement rate on it
best rate, different post
👋 Psst, flip me for morePost Performance
Audience
👋 Psst, flip me for moreAudience
🏢 Top company: Cognizant 😊 Okay, flip me backI love turning bold ideas into AI and data products that people genuinely enjoy using. I thrive on building great teams, tearing down silos, and shipping things that make a real difference for both the business and the people who use them.
Want to geek out over AI product strategy, LLM and agentic apps, ML-driven platforms, or enterprise SaaS? Or just feel like saying hi? I'd absolutely love that. 😊 Reach out anytime — my inbox is always open, and I can't wait to connect!
ProblemResearchers face information overload, with critical knowledge scattered across papers, trials, conferences, and competitor data.
ApproachBuilt a RAG platform (Langflow + Chroma + GPT-5-mini) that lets researchers chat with the literature and hear it back via TTS.
ResultResearch time cut by 70%, with sourced answers unified across 4+ literature types.
Researchers face overwhelming information overload, with critical knowledge scattered across papers, clinical trials, conferences, and competitor data that is slow and hard to synthesize. The challenge was to let researchers quickly query all these sources in one place and get sourced, real-time answers in accessible formats.
Built ScienceAI, a Gen AI platform that lets researchers chat with scientific literature — papers, clinical trials, conferences, and competitor data — in one place. A Langflow RAG pipeline with a Chroma vector database and GPT-5-mini streaming (on Next.js 16) delivers sourced, real-time answers, while ElevenLabs TTS converts insights into on-demand audio summaries and dual-voice podcasts for faster, more accessible research consumption.
ProblemHigh volumes of customer reviews were too slow and costly to read, prioritize, and respond to manually.
ApproachBuilt an LLM-powered classifier that predicts category and priority, and drafts sentiment-based responses automatically.
ResultReview handling automated end-to-end, cutting response time and lifting customer satisfaction.
Businesses receive high volumes of customer reviews that are slow and costly to read, prioritize, and respond to manually. The challenge was to automate review classification, prioritization, and response so teams could manage feedback at scale while improving customer satisfaction.
Built an LLM-powered solution that automates the classification and processing of customer reviews — predicting categories, assigning priority, suggesting actions, and generating sentiment-based responses. The solution streamlines review management, enhancing customer satisfaction and operational efficiency in handling feedback and inquiries.
ProblemSpam SMS messages posed a growing phishing and productivity risk with no automated way to filter them.
ApproachBuilt a text-classification model trained on labeled spam/ham messages to flag risky texts automatically.
ResultHigh-accuracy spam detection, reducing phishing exposure and improving user trust.
Businesses face growing security and productivity risks from spam SMS messages that can carry phishing and cyber-attack attempts. The challenge was to automatically classify incoming SMS texts as 'spam' or 'ham' to protect users and prevent attacks at scale.
Built a text-classification model that analyzes SMS messages labeled 'spam' and 'ham' to extract meaningful signals and predict whether a new message is spam. The solution automates spam detection, helping prevent phishing and cyber attacks before messages reach users.
ProblemHotels lose significant revenue to last-minute cancellations with little ability to see them coming.
ApproachBuilt a predictive model identifying the key drivers behind cancellations and flagging at-risk bookings early.
ResultAt-risk bookings predicted in advance, enabling more profitable refund and policy decisions.
Hotels and booking platforms lose significant revenue to last-minute cancellations with little ability to anticipate them in advance. The challenge was to identify the key drivers of cancellations and predict at-risk bookings early enough to shape profitable cancellation and refund policies.
Built a predictive model that identifies the factors most influencing booking cancellations and flags which bookings are likely to be cancelled in advance. The solution enables data-driven cancellation and refund policies, helping reduce revenue loss and improve operational planning.
ProblemCompanies struggle to gather and act on customer feedback fast enough to drive product decisions.
ApproachBuilt an AI platform that ingests, categorizes, and analyzes feedback to surface themes, sentiment, and priority issues automatically.
ResultFeedback analysis time cut by 80%, with actionable insights surfaced in real time.
In a competitive landscape, companies struggle to continuously gather, analyze, and act on customer feedback fast enough to drive innovation and growth. The challenge was to streamline feedback analysis so teams could surface actionable insights quickly and at scale.
Built an AI-powered customer feedback analysis solution that automatically ingests, categorizes, and analyzes feedback to surface themes, sentiment, and priority issues. The platform turns unstructured feedback into clear, actionable insights, enabling teams to continuously gather, analyze, and act on customer signals.