Headline

AI/ML Leader | Freelance AI Consultant | Financial Services | Ex-HDFC Bank | Agentic AI · GenAI · LLMs · Computer Vision · NLP

About

AI/ML practitioner with 6+ years of experience building and deploying intelligent systems at scale -> now independently consulting for financial services firms on their AI transformation journey. Most recently at HDFC Bank, I led the AI R&D function, managing a team of 15 Data Scientists and AI Engineers. I architected and shipped a document processing platform powered by Gemini 2.5 Flash, Google ADK, and LiteLLM -> delivering 20 document pipelines to production and 40+ to UAT, covering virtually every form and document type in the bank's workflow. I also built a multi-agent system that accelerated automobile loan processing end-to-end, and designed a computer vision system using DeepFace, FAISS, and GAN-based face enhancement to detect duplicate account fraud across millions of customer records. What I bring to the table isn't just technical depth -> it's the ability to identify where AI creates real business value in financial services, architect the right solution, build the team, and deliver to production. My current focus areas include Agentic AI and multi-agent orchestration, LLM-powered document intelligence and process automation, computer vision and fraud detection systems, and end-to-end ML pipeline design on platforms like Databricks. I work with investment firms and financial institutions as a freelance AI consultant - helping them move from AI exploration to production-grade systems that drive measurable outcomes. Open to consulting engagements, advisory roles, and conversations about what AI can realistically do for your business today.

Location

Pune District

AI stack

anthropic apis

claude code

gemini

hugging face

langgraph

perplexity

Work experience

AI Strategy & Engineering Lead (Independent) | Agents · LLMs · Computer Vision at Independent / Freelancer, Apr 2026 - Present · 5 mos, Bengaluru, Leading AI strategy and hands-on implementation for banking and investment firm, building intelligent systems across the AI stack from agentic workflows and LLM-powered applications to computer vision and NLP pipelines. Scope includes evaluating and deploying foundation models, designing multi-agent architectures, and translating business objectives in financial services into production-grade AI solutions. Focus areas: Agentic AI, Generative AI, Large Language Models, Computer Vision, NLP, and classical ML.

AI Lead at HDFC Bank, Aug 2021 - Apr 2026 · 4 yrs 9 mos, Pune District, - IDRS: Engineered a GenAI Pipeline leveraging Gemini 2.5 Flash and advanced Prompt Engineering (Self-Reflection/Consistency) to automate key information extraction from financial documents, pushing the pipeline for 13 document types into UAT. - Photo Deduplication: Developed a Deep Learning fraud detection system using ArcFace/FaceNet512 embeddings and FAISS nearest-neighbor search; achieved 97% precision and 85% recall in capturing duplicate photo accounts. - Face Cropping for Regulatory Compliance: Delivered a critical, high-volume solution using OpenCV and Azure Face Detection for extracting customer photos from OVDs, processing over 3 million images to meet urgent RBI compliance requirements. - Code & Infrastructure Optimization: Led a major performance overhaul of PySpark/Python processes, migrating Pandas to Polars and implementing parallel execution in Databricks; reduced a key job's execution time from 7 hours to 50 minutes and a Python script from 4 minutes to 57 seconds.

Senior Software Engineer / Data Scientist at Ness Digital Engineering, Aug 2018 - Aug 2021 · 3 yrs 1 mo, Pune, Maharashtra, India, > Credit Risk Analysis - Developed ensemble model for estimating the probability of defaults. > Market Risk Analysis - Created a Time Series model for estimating value at risk. > Customer Propensity - Developed predictive models for personal loan customer classification. > Optimized product dashboard queries to improve the load time by 90%. > POC - Demonstrated and Implemented a scalable data pipeline to efficiently ingest ~70m records of data in ~8 minutes

Senior Software Engineer at Persistent Systems, Aug 2015 - Jul 2018 · 3 yrs, Nagpur Area, India, > Improved the rendering time of admin UI components by more than 70% > Strengthened the security of the application by mitigating 70+ security defects. > Designed and developed a background job monitoring solution for preventing critical production issues. > Independently implemented the feature to allow the automatic provisioning of Active Directory users.

Education

Master of Science - MS, Machine learning and Artificial Intelligence, Liverpool John Moores University

Post Graduate Diploma, Artificial Intelligence/Machine Learning, International Institute of Information Technology Bangalore