Naman Kumar 2026
PDF

Naman Kumar

Senior Software Engineer with 6+ years building production AI systems — agents, context retrieval, evaluations, and infrastructure. Scaled the AI coding assistant Cody from idea to multi-million ARR across 250+ enterprises, and built autonomous agents at Railway that diagnose and fix production failures. M.S. in Machine Learning & AI.

Experience

Railway

Dec 2025 – Jun 2026

Senior Product Engineer

  • Designed, built, and launched an autonomous deployment diagnosis and error monitoring agent that analyzes runtime logs, failed deployments, code, and configuration across millions of services deployed on Railway.

  • Architected the reliable execution and evaluation infrastructure for Railway Agents using Temporal workflows, enabling autonomous code fixes and continuous quality improvements.

  • Shipped billing, usage metering, and rate limiting for Railway Agents using Stripe Meters, supporting 300K+ paid subscribers consuming ~$2M annualized in agent tokens.

  • Designed log fingerprinting for a ClickHouse-based runtime log ingestion pipeline processing 1M+ RPS, enabling scalable detection of new, regressed, and high-volume production errors.

Sourcegraph

Apr 2022 – Dec 2025

Senior Fullstack Engineer

  • Scaled Sourcegraph's Cody from prototype to multi-million ARR AI coding assistant serving 250+ enterprises and 15K+ daily developers including Netflix, Uber, Canva, Grab and Booking.com.

  • Designed Cody's hybrid multi-repo context retrieval pipeline using BM25, vector embeddings, and reciprocal rank fusion (RRF), enabling reasoning across thousands of repositories. Evaluated OpenAI and Sentence Transformer embeddings.

  • Implemented OpenCtx before the emergence of MCP, then later adopted the industry-standard Model Context Protocol (MCP) to enable Cody to retrieve context and execute actions across GitHub, Slack, Sentry, Datadog and other developer tools.

  • Built monetization infrastructure with Stripe subscriptions, usage-based rate limits, and abuse prevention, enabling Cody's PLG-driven growth to multi-million ARR.

  • Tech Lead for Cody Web, launching standalone Cody on the web, embedding inline AI assistance across code search, and shipping prompt libraries and a unified AI + code search experience across VS Code and JetBrains, powered by intent detection.

  • Tech Lead for Permission Center, rebuilding code-host permission sync across thousands of repositories and users per customer, and giving admins a faster way to audit repo access and sync status.

  • Built in-product analytics dashboards quantifying developer time saved and customer ROI, enabling major enterprise expansion deals.

Product Hunt

Sep 2019 – Mar 2022

Fullstack Engineer

  • Built, launched, and A/B tested 10+ product growth experiments, including personalized home feeds, content gating, promo codes, and ad placement optimizations, driving user acquisition and retention.

  • Built rule-based spam detection services that automatically flagged 80% of fake accounts, significantly reducing moderation efforts by the support team.

  • Improved page load times and cut origin load by building a Cloudflare CDN edge-worker caching layer that served cached pages to logged-out users while conditionally bypassing the cache for users in active A/B experiments.

  • Built and operated the platforms for Product Hunt's flagship community events, including the annual Golden Kitty Awards and quarterly Maker Fest hackathons, handling nomination and voting flows at peak traffic.

Internships

GraphQL Foundation

May 2020 – Aug 2020

Google Summer of Code

  • Developed a fast, maintainable streaming GraphQL parser adopted in graphql-js and GraphiQL, improving performance and enabling syntax-aware highlighting.

Grofers

May 2019 – Aug 2019

Fullstack Engineer Intern

  • Automated customer support at Grofers by building a self-serve chat system that handled ~12K daily tickets during peak sales, cutting manual workload by 65% and reducing mean response time by 25 minutes.

Redacumen

Jan 2019 – Aug 2019

Contract Fullstack Engineer (Part-time)

  • Built an accounts-payable automation platform that cut AP operations cost by up to 85%, including an OCR-based document aggregation and auto-indexing pipeline (email, scanner, cloud) and an ERP sync agent.

  • Built a cloud kitchen food-ordering platform spanning three apps: a consumer app for subscription-based ordering, a cook-facing app for menu and order management, and a delivery-hub app for dispatch coordination.

HackerEarth

Jun 2018 – Apr 2019

Frontend Engineer Intern

  • Led frontend development of FaceCode, a remote interview platform with real-time coding (WebSockets), live video, and code execution.

  • Built enterprise analytics, billing, and team management features, including RBAC, invitations, and admin workflows.

Coding Blocks

Jun 2017 – Nov 2018

Software Engineer Intern

  • Built an online IDE supporting 13 programming languages with real-time collaboration features.

  • Wrote a static site generator that cut code redundancy by 80% and saved hours of daily content-management effort for the website.

  • Taught Data Structures & Algorithms and React.js to 100+ students.

Education

Liverpool John Moores University

2022 – 2024

M.S. in Machine Learning & AI

  • Thesis: Optimizing Information Retrieval in Conversational Context via LLM Query Rewriting

Guru Gobind Singh Indraprastha University

2016 – 2020

B.Tech in Computer Science & Engineering

  • Project: Distributed Key-Value Fault-Tolerant In-Memory Database

Skills

  • AI Engineering: Coding Agents, AI Agents, RAG, MCP, Context Engineering, Browser Automation, Token Compression, LLM Evaluations, LLM-as-Judge, NLP

  • Search & Retrieval: Code Search, Hybrid Search, Information Retrieval System Evaluations, BM25, Vector Search, Embeddings, Sentence Transformers

  • Backend & Observability: Temporal, ClickHouse, Postgres, Kafka, Redis, Docker, Kubernetes, OpenTelemetry, Datadog, Sentry, AWS, GCP, Cloudflare, Vercel

  • Languages & Tools: TypeScript, Python, Go, Rust, SQL, React, Next.js, GraphQL, gRPC, tRPC, Playwright, Vitest, Ruby