Headline

AI Engineer | Building Production LLM Systems | Agentic AI, RAG & Applied AI | Python, AWS

About

Vedaant is an AI Engineer at TresVista Analytics, focused on building production-ready LLM applications, agentic AI systems, and intelligent automation workflows. He designs and develops AI solutions that combine large language models, machine learning and cloud infrastructure. His work includes multi-agent orchestration, RAG, structured data extraction, document AI, OCR, entity resolution, and AI-powered decision systems. At TresVista, Vedaant has: • Built multi-agent systems using AWS Bedrock, AgentCore, Claude, and LangGraph. • Developed a five-phase AI pipeline that extracts structured information from healthcare pitch decks using parallel People, Therapy, and Finance agents. • Built an AI-powered investment-memo generation workflow using parallel LLM agents. • Developed event-driven LLM workflows for email classification, document extraction, duplicate detection, company matching, CRM enrichment, and automated reporting. • Improved company-name reconciliation accuracy by more than 30% and eliminated more than 20% of duplicate records using LLM-based entity resolution. • Reduced manual deal-triage time by approximately 40% through AI-assisted automation. His technical interests include production LLM systems, agentic AI, RAG, multi-agent architectures, prompt engineering, model evaluation, document intelligence, applied AI, and scalable AWS infrastructure. Vedaant earned his B.E. in Electronics and Telecommunication Engineering from RV College of Engineering, Bengaluru. He is passionate about turning emerging AI capabilities into reliable, scalable products that solve real-world problems. For collaboration, AI engineering discussions, or new opportunities, Vedaant can be reached at vedaantmukherjee14@gmail.com.

Location

Bengaluru

AI stack

claude code

langchain

langgraph

langsmith

mcp

openai

Work experience

AI Engineer at Financial Services Company, Jul 2024 - Present · 2 yrs 2 mos, Bengaluru, Karnataka, India, • Built multi-agent AI systems for venture capital investment intelligence using AWS Bedrock, AgentCore, Claude, and LangGraph. • Developed a five-phase AI pipeline that extracts structured people, therapy, and finance data from healthcare venture capital pitch decks. • Built an AI-powered investment-memo workflow using parallel LLM agents to support venture capital deal evaluation. • Developed event-driven LLM workflows for inbound deal classification, document extraction, duplicate detection, company matching, CRM enrichment, follow-up automation, and investment-opportunity reporting. • Improved company-name reconciliation accuracy by more than 30% and eliminated over 20% of duplicate records in venture capital opportunity data through LLM-based entity resolution. • Reduced manual venture capital deal-triage time by approximately 40% through AI-assisted automation.

Co-Founder at Goffin, Nov 2021 - May 2022 · 7 mos, Bengaluru, Karnataka, India, Goffin was basically a platform where home-cooked meals would be available for bachelors who travel all the way from their native place to different cities for studies or jobs. For this the suppliers were the housewives and the consumers were the bachelors. We had good amount of positive responses from both sides regarding this idea. Unfortunately, we could not crack the delivery thing . This is where we failed and we had to shut down.

Education

Bachelor of Technology - BTech, Telecommunications Engineering, RV College of Engineering

Don Bosco School,Siliguri

Public links

github: https://github.com/vedmukherjee15