Headline

AI & ML Engineer | LLM/RAG · Agents (MCP, Claude Agent SDK) · Data Pipelines | Python · SQL | ex-Start.io, ex-EXL · 10+ yrs

About

Data & AI engineer with 10+ years across data analytics, machine learning, and data engineering — and, most recently, LLM, RAG, and agent systems. At Start.io I built audience and segmentation systems for a real-time mobile-advertising platform: app-text embedding models with XGBoost, Spark/OCI POI clustering, a RAG LLM page classifier, and Python/Jenkins data pipelines — progressing from Data Analyst to Senior Data Engineer. Earlier, over 5 years at EXL in London, I built statistical models that drove real money: customer-lifetime-value (GLM), churn and bad-debt models (logistic regression, ~82% lift, 3pp churn reduction), and pricing models worth £440k in value upside. Most recently, as co-founder of Appweave Labs, I've gone deep on applied AI: an LLM + deterministic-rules HS-code classification engine with a swappable multi-embedder RAG layer (pgvector) and an evaluation-first workflow; an internal multi-agent platform on the Claude Agent SDK with a hybrid RAG "Brain" exposed as an MCP server and an autonomous bug-to-PR coding agent; and production data/analytics tooling. I care most about the part that decides whether AI actually works: retrieval quality and honest evaluation. I've worn operator hats too — hiring, client delivery, GTM — but the work I go deepest on, and want to keep doing, is technical: data and AI systems you can trust. Stack: Python, SQL, R · LLMs, RAG, embeddings, pgvector, model evaluation · XGBoost, scikit-learn, Spark · LangChain, Claude Agent SDK, MCP · FastAPI, Flask, Next.js · AWS, Docker, PostgreSQL.

Location

Bengaluru

AI stack

claude code

claude

codex

langchain

obsidian

opencode

Work experience

Co-Founder, AI & Data Engineering at Appweave Labs, Feb 2024 - Present · 2 yrs 9 mos, Bengaluru, • Solo-built an LLM + deterministic-GRI HS-code customs-classification engine (TypeScript/Next.js, Postgres/pgvector) that abstains rather than guess, ~halving confidently-wrong filings vs. a single-shot LLM on benchmark rulings. • Architected an internal multi-agent platform (Python/FastAPI, Postgres/pgvector) with a hybrid RAG "Brain" (MCP server) and an autonomous bug-to-PR coding agent on the Claude Agent SDK. • Led architecture and POCs (as technical PM) for a watch-marketplace data platform's ingestion tooling — deterministic per-site scrapers (300+ tests), an SSRF-guarded "scrape-advisor" agent, and a semantic-search MCP server — built out by the team. • Built a founders/admin analytics dashboard (Flask + Next.js/recharts) aggregating GMV/MRR/AR/funnel KPIs from production Postgres across four analytics sources behind role-gated APIs. • Built the POC and core architecture for Certifico — tamper-proof credential verification via Solana compressed NFTs (Metaplex Bubblegum) with QR — productionised by my team. • As co-founder, hired a ~10-person eng/QA team, owned delivery for 4 clients (Collected, EXL, Medicle AI, HOA), and led GTM for an in-house 3D-printing product.

Senior Data Engineer at Start.io, Mar 2025 - Mar 2026 · 1 yr 1 mo, Bengaluru, • Implemented a RAG LLM to classify web pages into taxonomies and audiences. • Built age/gender prediction models from app-description text embeddings + XGBoost. • Built Spark/OCI clustering of Points-of-Interest visits and proximity audiences for top U.S. brands. • Built the JSON-to-SQL module for MAIA, an NLP/AI audience-building platform. • Established Python/Jenkins data pipelines for third-party ingestion; built internal Streamlit/Flask tools.

Data Engineer at Start.io, Jul 2023 - Feb 2025 · 1 yr 8 mos, Bengaluru, Karnataka, India

Data Analyst at Start.io, Mar 2021 - Jun 2023 · 2 yrs 4 mos, Bengaluru, Karnataka, India

Senior Consultant at EXL, Oct 2019 - Feb 2021 · 1 yr 5 mos, London, United Kingdom, Ownership of Bundles Proposition analytics, a new entry level product with free trial leading to auto renewal – look at engagement, product survival, operational issues etc. Mentored a team of 5 offshore analysts in various home services related analysis like drive to online, engineer visit optimisation, customer retention, product performance etc.

Consultant at EXL, Oct 2018 - Sep 2019 · 1 yr, London, United Kingdom, Built a model which estimates customer life time value by calculating discounted cash flow over next 5 years for various energy and services products. Used log linked gamma distributed GLM to model cost to serve, claims risk and logistic regression to predict churn and bad debt propensity – which collectively helped de-average cost on customer level

Senior Business Analyst at EXL, Oct 2017 - Sep 2018 · 1 yr, London, United Kingdom, Forecasting Sales and Market Share Built a linear regression model to forecast regional sales, market share and average consumption for a product being priced. Helped generate a value upside of £440k (H1 2018) by optimizing the acquisition price of online only energy products exclusive to price comparison websites Customer Retention Model Built a logistic regression model to predict customer propensity to churn at product renewal using customer demographics and price elasticity - obtained significant cumulative lift ~82% Implemented the model to pre-engage price sensitive customers with cheaper deals and realized churn reduction of 3pp

Summer Intern at DRDO, Bangalore, May 2014 - Jul 2014 · 3 mos, Bengaluru Area, India, Implementation of MATLAB Based GUI for Phased Array Radar System Calculation: Studied basics of radar engineering, developed an application and user interface for radar simulation.

Steering Committee Member at Avirbhav, University of Allahabad, Sep 2013 - Mar 2014 · 7 mos, Allahabad, Uttar Pradesh, India, Planning and execution of the college cultural fest attended by close to 500 people. Supervised a team of 60 organizing members, guided and kept them motivated. Projected the fest with an organic reach of 4000, and a viral reach of 12000 people.

Head of Design and Media at Avirbhav, University of Allahabad, Sep 2013 - Mar 2014 · 7 mos, Allahabad, Uttar Pradesh, India, Designed the branding, posters, website, video trailers for publicity and social media.

Education

Bachelor of Technology (B.Tech.), Electronics and Communication Engineering, University of Allahabad

Senior School, Mathematics and Computer Science, Modi Public School