# Ishan Avasthi

> AI Engineer — I build LLM-powered systems, agentic workflows, and developer tools.

I'm a CS undergrad at BITS Pilani working on AI engineering. I build the layer between raw language models and working products: agent pipelines, RAG systems, and the observability and evals that make them reliable enough to ship.

- **Role:** AI Engineer
- **Based in / studying at:** IND · BITS Pilani
- **Status:** open to AI engineering roles
- **Email:** hello@ishanavasthi.in
- **Website:** https://ishanavasthi.in
- **GitHub:** https://github.com/ishanavasthi
- **LinkedIn:** https://www.linkedin.com/in/ishanavasthi

## Docs

- [Full context](https://ishanavasthi.in/llms-full.txt): every project's full description, stack, links and commit activity, plus the complete text of every blog post, in one file
- [Projects page](https://ishanavasthi.in/projects): the filterable project archive
- [Sitemap](https://ishanavasthi.in/sitemap.xml)

## Skills

- **Agents & LLM Systems** (the core focus): Multi-agent pipelines, MCP, Tool calling, Claude API, Gemini, Groq, LangGraph, LangChain, Voice agents · LiveKit, RAG · ChromaDB — 15 projects
- **Evals & Reliability** (what makes agents shippable): Agent evals, OpenTelemetry tracing, LangSmith, RL environments · OpenEnv, CI regression gates, Deterministic guardrails, Prompt-injection defense, SigNoz / ClickHouse — 9 projects
- **Languages** (daily drivers first): Python, TypeScript, JavaScript, Java, SQL, C# — 14 Python, 13 TS
- **Backend & Infra**: FastAPI, Node / Express, Spring Boot, Postgres, Supabase, MongoDB, Docker, GitHub Actions, AWS, asyncio / TCP — 15 projects
- **Frontend**: Next.js, React, Tailwind CSS, React Native, Chrome Extensions · MV3 — 9 projects

## Projects (33)

### Agents & LLM Systems

- [Agent Store](https://github.com/ishanavasthi/agent-store): Merchant-side agentic commerce: messy captions in, buyable MCP catalog out.
- [Preflight](https://github.com/ishanavasthi/preflight): CI gate that catches AI agent regressions before they ship.
- [ClinicFlow](https://github.com/ishanavasthi/clinicflow): Real-time AI voice receptionist for clinics, with a live operator dashboard.
- [AgentGrid](https://github.com/ishanavasthi/agentgrid): Nine Gemini agents take an issue from bug report to merged PR, unattended.
- [AlphaDesk](https://github.com/ishanavasthi/alphadesk): Multi-tenant portfolio analyzer for Indian investors, narrated by agents that can't invent a number.
- [Swiggy Claw](https://github.com/ishanavasthi/swiggy-claw): Order food, groceries, and tables in plain English over Swiggy's MCP platform.
- [ReviewGrid](https://github.com/ishanavasthi/reviewgrid): Agent-callable code review skill returning a deterministic verdict.
- [Panorama](https://github.com/ishanavasthi/panorama/tree/v2): Reviews a pull request using evidence from the other repos in the org.
- [PR Review Bot](https://github.com/ishanavasthi/pr-review-bot): GitHub App that reviews every pull request with Gemini.
- [Message Notification Router](https://github.com/ishanavasthi/personalized-message-router): Routes every WhatsApp message to notify, digest, or mute — and reproduces its own results offline.
- [Agent-Toolkit Buildability Atlas](https://github.com/ishanavasthi/composio-research): An agent pipeline researched 100 apps to answer one question: can we build a toolkit for it?
- [Swiggy Agent (n8n)](https://github.com/ishanavasthi/swiggy-n8n): Agentic food-ordering concierge built entirely out of n8n workflows.

### RAG & Document AI

- [Lenny Growth Assistant](https://github.com/ishanavasthi/lenny-growth-assistant): Grounded answers over 301 Lenny's Podcast episodes, cited to the exact YouTube second.
- [ChatWithPDF](https://github.com/ishanavasthi/ChatWithPDF): NotebookLM-style agentic assistant for querying your own PDFs.
- [LexGuard](https://github.com/ishanavasthi/lexguard): Upload a contract, get plain-English risk analysis of every clause.

### RL Environments & Evals

- [RecruitEnv](https://github.com/ishanavasthi/recruit-env): OpenEnv RL environment for candidate pipeline triage under a step budget.
- [InsuranceClaimEnv](https://github.com/ishanavasthi/insurance-claim-env): OpenEnv environment for high-stakes insurance claim adjudication.

### Developer Tools

- [ContextLens](https://github.com/ishanavasthi/context-lens): Browser activity capture that survives MV3 service worker death without dropping or duplicating a single event.
- [Markdown Copier](https://github.com/ishanavasthi/markdown-copier): Chrome extension that turns any page into clean, LLM-ready Markdown.
- [KV-Cache](https://github.com/ishanavasthi/kv-cache): In-memory key-value store over raw TCP, built on asyncio.

### Full-Stack & Web

- [Verdict Desk](https://github.com/ishanavasthi/verdict-desk): Grades student code in a hardened Docker sandbox, and no AI answer reaches a student until a teacher approves it.
- [FinTracker](https://github.com/ishanavasthi/fintracker): Personal finance tracker with category budgets and live spend breakdowns.
- [EventHive](https://github.com/ishanavasthi/EventHive): Event discovery and booking platform with a React Native client.
- [This Portfolio](https://github.com/ishanavasthi/portfolio): The site you're reading — Next.js, MDX blog, and this cascade.

### Foundations

- [Product Service](https://github.com/ishanavasthi/Product-Service): Spring Boot product catalogue microservice with Flyway migrations.
- [Cart Service](https://github.com/ishanavasthi/cart-service): Spring Boot cart service with layered domain design.
- [Patient Record System](https://github.com/ishanavasthi/Patient-Record-System): Spring Boot REST API for patients, histories, and appointments.
- [Book Management](https://github.com/ishanavasthi/spring-book): Spring Boot book and author CRUD with a full test suite.
- [Bestsellers Analyser](https://github.com/ishanavasthi/lld-assignment): Java CSV analysis of Amazon's top 50 bestsellers, 2009–2019.
- [Image Editor](https://github.com/ishanavasthi/image-editor): Image manipulation in pure Java — grayscale, blur, rotate.
- [Aerofit Sales Analysis](https://github.com/ishanavasthi/aerofit-analysis): Exploratory case study on Aerofit treadmill sales.
- [Easy App](https://github.com/ishanavasthi/rn-assignment-ishanavasthi): React Native unit converter with live conversion.
- [Overloading](https://github.com/ishanavasthi/OverloadingMVP): Unity game-jam prototype about not tipping over a loaded truck.

## Writing

- [Your agent got 3x more expensive and every test still passed](https://ishanavasthi.in/blog/catching-agent-regressions-in-ci): Building a CI gate for AI agents, with SigNoz as the datastore rather than a dashboard on the side. Every bug that mattered was silent, and five separate green lights meant nothing. (2026-07-27)

## Optional

- [GitHub](https://github.com/ishanavasthi)
- [LinkedIn](https://www.linkedin.com/in/ishanavasthi)
- [Email](mailto:hello@ishanavasthi.in)