Available for AI architecture & consulting
I architect AI-native products that ship.
Senior AI Architect & Full-Stack Product Engineer with 12+ Years. Agentic AI, RAG pipelines, and scalable full-stack platforms — from concept to production.
I architect Agentic AI ecosystems, RAG pipelines, and scalable distributed systems — then ship them as production-grade products. Over 12+ years I've led cross-functional teams across healthtech, fintech, e-commerce, and HR-tech, migrating legacy monoliths to microservices and delivering AI-native platforms that move real business metrics.
How I can help
AI consulting · Agentic systems · LLM engineeringAI Consulting & Strategy
From use-case discovery to architecture and ROI modelling — I help teams turn AI ambition into a shippable, measurable roadmap.
- Use-case discovery
- AI roadmaps
- Build vs. buy
- Cost & latency modelling
Agentic AI Systems
Autonomous, multi-step agents with reasoning, tool use, and self-reflection loops — orchestrated for durability and reliability.
- LangGraph orchestration
- Multi-agent workflows
- Temporal durability
- Human-in-the-loop
RAG & LLM Engineering
Production RAG pipelines with rigorous evaluation — faithfulness, precision, and continuous benchmarking baked in from day one.
- Vector pipelines
- RAGAs evaluation
- LangSmith observability
- Hallucination control
Full-Stack Product Engineering
End-to-end delivery — web, mobile, and backend — as monorepo platforms, from architecture to App Store and production.
- Next.js + FastAPI/NestJS
- React Native apps
- Nx / Turbo monorepos
- CI/CD & cloud
Industries I've shipped in
Technical expertise
The modern AI & full-stack toolkitAI & Agentic Systems
- LangGraph
- LangChain
- Autonomous agents
- Multi-step reasoning
- Prompt engineering
- Self-reflection loops
LLM Evaluation & Observability
- RAGAs (faithfulness, precision)
- LangSmith tracing
- Evals & benchmarking
- Continuous quality gates
Vector & Data Infrastructure
- Pinecone
- Weaviate
- PostgreSQL
- MongoDB
- Redis
- Prisma ORM
Systems Architecture
- Microservices (NestJS, Moleculer)
- Event-driven (Kafka, RabbitMQ)
- Temporal orchestration
- Azure Service Bus
Full-Stack & Mobile
- Next.js
- FastAPI (Python)
- NestJS
- React / React Native
- TypeScript
- GraphQL
Monorepo & DevOps
- Nx / Turbo monorepos
- Docker & Kubernetes
- CI/CD (GitHub Actions, Azure)
- AWS & Azure
Selected work
Client names are protected under NDA. What follows is the scope, architecture, and technology behind the products.
Building AI-native SaaS products across healthtech, Ayurveda, and workforce-management verticals.
- Architected a cloud-based physiotherapy EMR OS (web + iOS + Android) from the ground up as an Nx monorepo with shared TypeScript logic across three surfaces.
- Engineered the AI feature layer: Speech-to-Notes SOAP auto-fill (~45 min/day saved per clinician), AI exercise coach with rep counting, and camera-based motion tracking.
- Built a RAGAs evaluation framework for an Ayurvedic clinical-intelligence platform, monitoring retrieval precision and answer faithfulness via LangSmith dashboards.
- Designed an LLM-driven active-recall engine for an EdTech platform — auto-generating spaced-repetition prompts with adaptive difficulty.
Led an engineering team of 10+ across an AI-native ATS and a nonprofit fundraising SaaS platform.
- Spearheaded an AI-native recruiting agent that parses, semantically matches, and ranks candidates using prompt engineering and LLM self-reflection loops.
- Integrated ElevenLabs voice synthesis + Daily.co WebRTC for human-like AI voice interviews — cutting time-to-hire by 40–60%.
- Implemented Temporal for durable execution of long-running, multi-day recruiting workflows with zero data loss from partial failures.
- Migrated a Node.js monolith to a Moleculer microservices architecture; standardized LangSmith observability across the department.
Led end-to-end delivery across e-commerce, fintech, education, and enterprise verticals.
- Re-architected a 200,000+ user enterprise member portal from a Magento monolith to a modern MERN + headless-CMS stack with sub-2s page loads.
- Delivered India's first B2B online appliance marketplace from concept to ₹100 Cr GMV in its first operating year.
- Built a full-cycle fintech LOS/LMS with automated credit decisioning (CIBIL/Bureau APIs) and DigiLocker / Aadhaar eKYC workflows.
- Scaled a sports-academy CRM 5X in revenue through product-led retention automation.
Forex/Crypto CRM delivery, on-demand delivery platforms, and premium e-commerce engineering.
- Managed delivery of a Forex broker CRM with 350+ configurable parameters integrated with MetaTrader 4/5 and cTrader, plus a 40+ crypto deposit extension.
- Built Talabat-style on-demand delivery platforms with real-time order tracking, multi-vendor management, and rider dispatch engines.
- Developed a proprietary Python size-matching algorithm mapping 20+ body measurements to garment specs, reducing fashion-return rates.
Impact in numbers
GMV delivered — India's first B2B appliance marketplace, year one.
Reduction in time-to-hire via AI voice interview automation.
Saved per clinician through AI Speech-to-Notes SOAP automation.
Revenue scale via custom CRM and product-led retention engineering.
Users migrated to a modern MERN stack with no downtime.
Drop in production incidents from PR quality gates & observability.
How I work
Discover
Map the problem, constraints, and the metric that matters. No solution before the why is clear.
Architect
Design the system — data, models, evals, and infra — for scale, cost, and reliability.
Build
Ship in tight iterations with quality gates, observability, and tests from the first commit.
Scale
Harden, monitor, and optimize — turning a working product into a durable platform.
What collaborators say
He doesn't just build features — he architects systems. Our AI hiring funnel went from idea to production with reliability we didn't think was possible at our stage.
Took our legacy monolith and a 200K+ user base and migrated it without a single day of downtime. Rare combination of architecture depth and delivery discipline.
The RAG evaluation framework he set up cut our hallucinations dramatically. He treats AI quality like an engineering problem, not a guessing game.
Let's build something intelligent.
Available for AI architecture, fractional engineering leadership, and product builds. Based in Noida, India — working globally.