Senior AI Engineer | Agentic AI, LLMs & RAG | Production systems on GCP | 12 years shipping AI that scales beyond demos
I'm a Senior AI Engineer with 11+ years of experience designing and shipping production AI systems, currently focused on agentic AI, large language models, and retrieval-augmented generation.
For the past several years I've worked at Oracle (formerly Cerner), building AI systems that operate at scale in regulated, real-world environments — not demos. My two flagship builds are an agentic claims processing system handling 2.8M+ healthcare claims using LangGraph multi-agent orchestration on GCP, and a Clinical Trial Navigator using RAG with PubMedBERT embeddings and Pinecone that improved precision from ~40% to 85%+.
What I bring to client work:
Agentic AI architecture using LangGraph, MCP (Model Context Protocol), and A2A protocols
LLM application development with Anthropic Claude, OpenAI, Llama, and Groq
RAG pipeline design with vector databases, hybrid retrieval, and evaluation frameworks
Fine-tuning and PEFT/LoRA for domain-specific tasks (33% better than GPT-4o on a healthcare pricing task)
Production deployment on GCP — Cloud Run, BigQuery, GCS, Secret Manager, Firestore, Cloud Monitoring
LLM observability, eval harnesses, and multi-tenant SaaS isolation
Recent measurable outcomes:
67% reduction in manual screening effort via agentic workflows
40% reduction in claims processing time
28% improvement in clinical data extraction accuracy
65% lower error than classical ML baselines
I work best on greenfield AI builds, taking LLM POCs into production, agentic system design, and RAG architecture reviews. Available for contract and consulting engagements with US and European clients. I prefer hands-on technical work over pure advisory.
If you're trying to get an AI feature past the prototype stage and into something reliable enough to put real users on, I can help.
Work Terms
Availability
20–40 hours per week for active engagements
Available for both short-term sprints (architecture reviews, POCs) and long-term builds
Bangalore, India (IST). Comfortable with significant overlap for US East Coast and all European time zones.
Engagement Models
Hourly contract work
Fixed-scope project engagements
Retainer / fractional AI engineering
Architecture review and advisory
Communication
Slack or your team's preferred messaging channel — responsive within a few hours during working hours
Weekly sync calls + async written updates by default
Zoom / Google Meet / Microsoft Teams all fine
Payment Terms
All payments through Guru SafePay
Milestone-based for fixed-scope projects
Weekly or bi-weekly invoicing for hourly engagements
USD or EUR preferred
What I need from you
Clear problem statement and the ability to share relevant data and systems
A point of contact who can answer domain questions
Defined success criteria — I work best when "done" is measurable