Open to software engineering roles

Palguna Shetty.

Software engineer · Karnataka, India

I build memory systems for AI agents and tools for developers. Interested in what agents remember, how they use it, and how to measure whether it helps.

About

From idea to working system

I build AI infrastructure, backend systems, and developer tools from 0 → 1.

Recently, I built an AI memory and context engine for conversational agents, covering ingestion, extraction, reconciliation, retrieval, and context generation. I benchmarked its retrieval system on 500 LongMemEval-S questions, reaching 95.4% Recall@5 and 98.4% Recall@15.

I also led development of Novel Signal, a competitive intelligence system for a client with over $10 million in revenue. It tracks competitors in real time across Amazon, Flipkart, and Meesho, including keyword rankings.

My work spans AI agents, LLM infrastructure, retrieval, backend engineering, and developer tools. I like owning the full engineering loop: understanding the problem, designing the architecture, building, benchmarking, shipping, and fixing what breaks.

Outside of coding, I led my department’s basketball team to a silver medal.

Retrieval benchmark
95.4%Recall@598.4%Recall@15 500 LongMemEval-S questions · my benchmark run
Looking for
Software Engineer, AI Engineer, and Founding Engineer roles at ambitious startups.
Based in
Karnataka, India
Education
B.E. Computer Science
JSS Science and Technology University
2022–2026
Languages
English, Kannada, Hindi

Selected projects

Code & experiments

OpenMem

Agent memory
Python · SQLite · LongMemEval

An embedded memory engine that stores conversation events, extracts durable memories with evidence, reconciles updates, and retrieves useful context.

FreshBrief

Research agent
Python · LangGraph · LangChain

Finds relevant information and turns it into concise research briefs with source citations.

Revflo

AI code review
React · FastAPI · Supabase · Groq

Reviews pull requests and repositories to surface useful code-quality feedback for engineering teams.

EvalSmith

Open source · MIT
Python · Pandas · LLM APIs · CI

Runs prompt regression tests to catch changes in LLM output quality before production.

Skills

What I work with

Languages

TypeScript, Python, C++, SQL, Go

Web & Backend

Node.js, FastAPI, Next.js, React

AI Systems

LangChain, LangGraph, LLM APIs, MCP

Data & Infrastructure

PostgreSQL, SQLite, MongoDB, Docker, AWS, Vercel, Render

Contributions

On GitHub

@palguna26

Public contributions over the last year

GitHub Profile

Loading contributions…

Writing

Notes from building

AI systems · 8 min read

Memory systems for AI agents

How memory works, why agents need it, and the practical decisions behind building a reliable memory layer.

Get in touch

For engineering opportunities, open-source projects, or a conversation about what you’re building.