Specialising in Gen AI, LLMs, and RAG pipelines that solve real-world financial problems — from automated loan decisioning to document intelligence and workforce risk analytics.
Production-grade 7-agent LangGraph system automating end-to-end mortgage underwriting — document intake, ratio analysis, policy compliance, risk scoring, and decisioning. Hybrid RAG retrieval with BM25, pgvector, RRF fusion, and cross-encoder reranking. AWS Bedrock, LangSmith observability, full audit trail.
Three-agent LangGraph pipeline automating vehicle service case triage, RAG-powered technical summarisation, and async Salesforce writeback via SQS — with real-time Slack alerting for critical incidents and zero manual intervention.
RAG-powered system for extracting and answering questions from complex financial documents — annual reports, contracts, regulatory filings — with citation-level accuracy and hybrid semantic plus keyword retrieval.
Conversational financial assistant leveraging LLMs with tool-use for real-time market data, portfolio analysis, and personalised guidance — with explainable reasoning and compliance guardrails.
AI-driven risk assessment system combining predictive modelling, anomaly detection, and automated reporting for enterprise workforce and financial obligation risk management.
I'm a software engineer specialising in Generative AI and financial systems — bridging the gap between cutting-edge LLM research and production-grade applications that solve real business problems.
My work spans multi-agent orchestration, RAG pipelines, document intelligence, and risk analytics — always focused on accuracy, auditability, and real-world impact.
Deep dive into the technical decisions, architecture evolution, problem-solving process, and lessons learned across all four AI projects — told chronologically.
Open to Senior GenAI/RAG Engineer roles in India. Reach out via email, LinkedIn or WhatsApp.