Real problems. Real systems. Shipping continuously.
Client engagements and capability demonstrations. Every project either replaced manual work with a system, or replaced a demo with something that runs in production.
VK Bags
From Zero Digital Presence to Professional Platform
A manufacturer supplying to multiple gyms including Gold's Gym, Nitro Fitness, Anytime Fitness and many more across India had no website. Every new client relationship started with cold trust and no digital reference point.
A full website with product catalogue, client logos, and a contact system - built and live in weeks. A manufacturer who previously existed only through word of mouth now has something to send a prospect before the first call.
Nutrition CRM
Client Management Platform for Coaches
A nutrition coach spent hours every day managing clients manually and diet plans in a diary, macros calculated by hand, appointments tracked in spreadsheets. There was no system. Just accumulated manual effort.
Everything that used to live in a diary, a spreadsheet, and a calculator now lives in one place. The coach spends less time on admin and more time actually coaching. Clients get their own portal to follow their progress.
SaaS products I'm building, owned, not client work.
SchemaHealer
AI-Powered CRM Data Migration & Recovery
CRM and legacy data often arrive with renamed columns, missing values, duplicate records, and inconsistent formats. That breaks imports, slows migrations, and creates expensive cleanup work before teams can trust the data.
SchemaHealer profiles messy source data, maps fields to the target schema, flags duplicates and format issues, and produces validated, import-ready output with a human review step before repair and export.
DocuLens
Hybrid SQL + Vector Search
Business data lives in two places that never talk to each other that is a database full of structured records and a folder full of documents. Answering a question that spans both means switching tools and piecing it together manually.
DocuLens routes every natural language query to the right engine automatically - SQL for structured data, vector search for documents, or both in parallel. One interface for everything the business knows, regardless of where it lives.
Proof-of-concept builds that show what's possible.
FilmOracle
Hybrid Recommendation Engine
Picking something to watch or play across five platforms takes longer than it should. Most recommendation engines suggest what's popular and not what fits the person asking.
Context-aware recommendation engine combining TF-IDF, semantic search, and LLM reasoning with confidence scoring across a unified movie and game catalogue. Tells you why it's recommending something, not just what.
Recovia
ML-Powered Loan Risk System
Lenders using manual rules to flag high-risk borrowers were always reacting too late chasing recovery after the default, not before it.
ML-based default risk prediction with SHAP explainability so every score can be justified, KMeans segmentation for recovery prioritisation, and automated PDF risk reports, served via a FastAPI and React dashboard.
Have a similar problem? Let's talk.
Book a free 30-60 min discovery call. You'll get a fixed-price proposal within 5 business days.