Available for full-time & internship roles

Kesav Patneedi

AI / ML Engineer

I'm Kesav, a final year Data Science and AI student at IIT Bhilai with 1.5 years of experience building production AI systems across startups and Fortune 500 companies. I work across the full stack from LLMs and agents to backend APIs and production deployments. I enjoy working in fast moving environments where I can own things end to end.

Get in touch View résumé
Portrait of Kesav Patneedi
91%
diagnostic accuracy from an automated clinical interview Valuai.io
9% → 1.2%
hallucinated claims in AI-generated clinical outputs Valuai.io
132M
transactions analysed for protection-plan attach rates Assurant
Top 50
of 32,000+ teams in the Amazon ML Challenge 2026

Where I've worked

Experience

  1. Jul 2025 — Present US · Remote Current role

    Valuai.io

    AI Engineer Intern

    • 91%diagnostic accuracy
    • 9% → 1.2%hallucinated claims
    • 62% → 86%plan-scoring accuracy
    • Built and deployed an AI-driven, patient-facing medical application from scratch, integrating voice agents, AI avatars and LLMs.
    • Automated a structured clinical assessment into a live, conversational interview, replicating clinician-grade branching logic using LangGraph and achieving 91% diagnostic accuracy.
    • Reduced hallucinations in AI-generated clinical outputs from 9% to 1.2% of claims using a citation-grounded verification loop with iterative regeneration.
    • Fine-tuned open-source models on real doctor review data to score AI-generated treatment plans and flag issues, improving scoring accuracy from 62% to 86% and feedback text similarity from 55% to 74%.
    • LangGraph
    • Voice Agents
    • Fine-tuning
    • RAG
    • FastAPI
    • React
  2. May 2026 — Jul 2026 India · Remote

    Assurant

    Data Science & AI Intern

    • 4,700+client stores served
    • 132Mtransactions analysed
    • 140K+graph relationships
    • Developed a full-stack internal analytics platform across 4,700+ client stores, generating store-specific natural-language insights and recommendations for client employees.
    • Conducted end-to-end driver analysis over 132M transactions to identify the key factors influencing protection-plan attach rates.
    • Designed and deployed a Neo4j knowledge graph with 140K+ store–driver relationships, enabling real-time store-level performance queries.
    • LLMs
    • Neo4j
    • Knowledge Graphs
    • Python
    • Analytics

Things I've built

Projects

Agentic Financial Document Retrieval System

An AI system that autonomously finds and analyzes financial documents to answer company-specific queries.

  • Built a supervisor–subagent architecture to autonomously discover, download and index SEC filings, annual reports and financial documents.
  • Applied Astute RAG for iterative, source-aware knowledge consolidation, resolving conflicts across multiple retrieved documents for reliable responses.
  • Implemented multi-hop query decomposition and financial jargon expansion to improve retrieval accuracy.
  • Multi-Agent
  • Astute RAG
  • LangGraph
  • Vector Search
View code on GitHub

Multi-Agent Trading & Portfolio Management System

An autonomous trading system that makes portfolio decisions using collaborative AI agents.

  • Built a real-time data pipeline using Redis Streams to ingest market, news, social and SEC signals.
  • Designed a multi-agent architecture that debates FinRL PPO model signals to make trade decisions.
  • Implemented automated trade execution and portfolio tracking using the Alpaca API.
  • Multi-Agent
  • Reinforcement Learning
  • FinRL
  • Redis Streams
  • Alpaca
View code on GitHub

Secure Implementation of Cryptographic Ciphers using Deep Neural Networks

Block ciphers implemented as neural networks, secured against key retrieval attacks.

  • Implemented PRESENT-80 and AES block ciphers entirely as PyTorch neural networks with fixed deterministic weights, encoding all cryptographic operations as custom NN layers.
  • Designed XORNet (ReLU-based), SBoxLayer (corner-function detection) and PermutationLayer (sparse matrix) as non-trainable PyTorch modules replicating exact cipher operations.
  • Hardened the implementations against key retrieval attacks, preserving cryptographic security properties within the neural network framework.
  • Verified correctness against all official test vectors with full pytest coverage across S-box, P-layer, key schedule and round key generation.
  • PyTorch
  • Cryptography
  • PRESENT-80
  • AES
  • pytest
View code on GitHub

Background

About

Achievements

Top 50

Amazon ML Challenge 2026

Ranked among the top 50 teams out of 32,000+ participating teams.

Lead

Data Science and AI Club, IIT Bhilai

Led the DSAI club in AY 2025 – 2026, organizing hackathons and knowledge-sharing sessions for students.

Education

Indian Institute of Technology Bhilai

B.Tech in Data Science and Artificial Intelligence

Aug 2023 — May 2027

8.84CGPA / 10

Technical Skills

AI & LLM
  • Python
  • LangChain
  • LangGraph
  • RAG
  • Hugging Face
  • PyTorch
  • AI Agents
  • Unsloth
  • QLoRA
  • STT
  • TTS
  • Sarvam AI
Full Stack
  • FastAPI
  • React
  • MongoDB
  • Redis
  • Neo4j
  • Supabase
Deployment
  • AWS S3
  • Bedrock
  • EC2
  • RDS
  • Cognito
  • SNS
  • Lambda
  • Docker
  • Caddy

Contact

Get in touch

I'm looking for full-time and internship roles. Email is the best way to reach me.

kesavpatneedi.work@gmail.com
Connect on LinkedIn GitHub View résumé