Flagship products, interactive demos, personal tools, and product concepts. Client and employer work is shown in summary; demos are clean-room reconstructions on synthetic data.
Batch-level traceability for global seafood exporters.
Owned the full lifecycle (vision, strategy, roadmap) for the SeaTrace AI MVP across Desktop, Tablet, and Mobile-responsive platforms
Score a medical claim's denial risk before it is ever submitted.
Started in the data, not the product: 8 commented SQL queries over a 6,000-claim synthetic warehouse established the leak was material, concentrated, preventable (80% of denials traceable to pre-submission causes), and not self-healing
Mine task events to show a practice where its staff hours actually go.
Task-event mining that ranks each practice's operational bottlenecks and sizes the automation ROI of fixing them
An LLM strategy agent where the model never computes a number.
Won a firm-wide AI & Sustainability innovation challenge as architect of the original concept
384 metrics, GRI/ESRS-aligned, reporting time cut by 25%.
Designed an environmental data model automating Scope 1–3 carbon tracking across 384 metrics aligned to GRI and ESRS, cutting reporting time by 25%
Data products for a $10tn+ AUM investment manager (de-identified).
Built 10+ AI-driven Tableau dashboards; automated PySpark/Python pipelines saving 100+ analyst hours/month
0→1 ESG product · Siemens Tech4Sustainability runner-up.
Runner-up among 152 global entries (USD 4,000, 'Tokenise the Energy Transition')
Manage Google Calendar by chatting, across cloud, local, and voice.
A natural-language agent with a six-tool calendar toolkit and streaming responses
A timeline for where your day actually goes.
Treats time as a continuous stream: every minute is Focus / Distraction / Neutral, one tap to switch
Focus-TimeLog, evolved: your day auto-ingested into one timeline.
Evolves Focus-TimeLog from manual tracking to automatic capture, with calendar, browser history, and health signals feeding one timeline
A second brain that answers: what did I learn, when, at what depth?
Capture and retrieval experiments for a personal knowledge system, built for myself first
What if advice found the investor, instead of the other way around?
A governed personalisation layer that detects real signals (job changes, accumulating cash, approaching retirement) and surfaces a contextual nudge, or deliberately stays quiet when the benefit is below threshold
Blockchain traceability for sustainable aviation fuel.
Organised a speaker session at the WEF Annual Summit 2023 on 'Achieving Sustainability through Supply-chain Traceability'
Incentivising flexibility in microgrids via blockchain.
A blockchain mechanism to incentivise flexibility & coordination in microgrids
India's first Web3 mega-event, 300+ artists.
Built a $250K+ sponsorship portfolio onboarding global Web3 brands
9,917 words merged from 20 sources, ranked by source quality rather than by how many lists happen to carry them.
Merged 20 sources (Barron's 4,841, GregMat's 37 groups, Magoosh, Manhattan, PowerScore, PrepScholar and three personal PDFs) into one deduplicated 9,917-word corpus, every word with a definition