Headline

AI Engineer | Turning Tokens into Impact | LLMs, Agents, Infra & Beyond

About

I build AI systems that ship to production and work at scale. Over the past 4 years, I've gone from building analytics dashboards to architecting multi-agent platforms that process millions of documents, generate clinical summaries, and automate complex business workflows. What I do best: → Multi-Agent Systems: Built agentic AI platforms processing surrogacy datasets into structured clinical summaries at Telomere. Designed multi-agent pipelines for lead research, tone control, and qualification at Accacia (10K+ users onboarded). → Document Intelligence: Architected a LayoutLM-based extraction system at Turtlemint — 95% F1 across 103 entity classes, 10M+ policy documents processed, latency reduced from 11 minutes to 11 seconds. → RAG & Retrieval: Built contract-aware RAG systems for compliance validation, GraphRAG-powered pitch deck generators using Neo4j, and hybrid search pipelines with re-ranking and context optimization. → Production Engineering: I don't build demos. Every system I've built handles real traffic — 10K+ requests/hour, Kafka/RabbitMQ message queues, AWS infrastructure with monitoring and alerting. Currently: AI Engineer at Telomere, building agentic AI for healthcare — turning fragmented medical records into structured clinical summaries that save medical teams hours of manual chart review. Previously: Shipped AI at Accacia (B2B outreach automation) and Turtlemint (document AI for insurance). Tech I work with daily: Python, FastAPI, LangChain, LangGraph, OpenAI, Claude, Neo4j, PostgreSQL, AWS, Docker, Kubernetes. Let's connect if you're building something interesting with AI, or if you need someone who can take your AI from prototype to production. 📅 Book a call: https://cal.com/anmol/intro 🌐 Portfolio: https://anmoljain.me

Location

Pune Division

AI stack

claude

codex

cursor

Work experience

AI Engineer at TELOMERE, Nov 2025 - Present · 9 mos, Washington, United States, Building an agentic AI platform for healthcare , processing large-scale surrogacy datasets and generating structured, citation-backed clinical summaries. Key work: • Designed multi-stage pipelines for ingestion, parsing, embedding, retrieval, and summarization across fragmented medical records • Implemented multi-agent workflows to extract pregnancy timelines, family medical history, prenatal/delivery summaries, and vaccination records • Generated structured 4-page clinical summaries, significantly reducing manual chart review effort for medical teams • Developed a contract-aware RAG system to validate surrogate expense requests against complex legal agreements • Automated DR request workflows with real-time compliance checks, reducing manual intervention • Deployed scalable infrastructure using AWS CDK and CloudFormation with load balancing, monitoring, and alerting Impact: Reduced hours of manual medical chart review to minutes. Enabled real-time compliance validation for surrogate expense requests.

AI Engineer at Accacia, Jan 2025 - Oct 2025 · 10 mos, Bengaluru, Designed and deployed an AI-driven outreach automation platform for hyper-personalized B2B messaging. Onboarded 10K+ users. Key work: • Built full-stack architecture (FastAPI, Supabase, PostgreSQL) with scalable cloud-native deployment • Implemented multi-agent pipelines for lead research, tone control, and qualification logic • Applied entity extraction + semantic scoring for dynamic personalization across company, industry, and persona context • Engineered an AI-powered Chrome extension automating LinkedIn and Gmail outreach with real-time, context-aware message generation • Built a pitch deck generation system using Neo4j + GraphRAG — modeled company/industry/persona relationships in graph databases and generated tailored PowerPoint sales decks Impact: 10K+ users onboarded with analytics via PostHog. Production-grade pipelines with polling, chunking, and fallback mechanisms.

Software Engineer at Turtlemint, Jul 2022 - Nov 2024 · 2 yrs 5 mos, •Engineered a robust end-to-end data extraction pipeline utilizing CLAUDE Haiku LLM, designed to process and extract complex data from PDF documents across diverse formats • Developed a flexible and configurable framework to efficiently store Prompts, Extraction keys, and Key Types, enabling seamless adaptability to extract data from any type of document • Innovated a custom in-house PDF policy parser for md conversion with enhancements to libraries like pymupdf and pdfminer, drastically optimizing token usage and boosting extraction accuracy to nearly 99%, integrating data from 40 insurers across motor, life, and health verticals • Led the architectural design and implementation of 6 microservices (Classification, OCR, NER, Backend, Frontend,QIS), achieving a dramatic reduction in data extraction time from 11 minutes to 11 seconds for a 80-page document for over 60 fields • Utilized SVM-based text classification with keyword analysis for page importance. • Used AWS Textract OCR service to extract text from important pages. • Trained LayoutLM Transformer-based NER model achieving over 92% accuracy. • Deployed model using Torch Serve and contributed to automated model training and deployment platform with AWS SageMaker pipeline integration • Implemented RabbitMQ and Kafka for efficient message queuing and processing, successfully handled over 1 million documents

Tech Intern at Turtlemint, Feb 2022 - Jul 2022 · 6 mos

Education

Bachelor of Technology - BTech, Computer Engineering, Vishwakarma Institute of Technology