Photo of Mrutyunjay Patil

Hi, Mrutyunjay here.

Applied AI Engineer

I build AI platforms from scratch. Expert in MCP and love SWE work.

To know more, have a quick chat with me

  • I

    Infosys Ltd.

    Specialist Programmer (Applied AI Engineer)

    • Built a multi-tenant agentic AI platform end to end and mostly solo: a FastAPI and LangGraph runtime, RAG knowledge bases, guardrails, and the Next.js studio UI on top. Other product teams now build their agents on it instead of writing their own runtime
    • Shipped the MCP infrastructure that exposes enterprise commerce to AI agents: a TypeScript MCP server for full B2B and B2C shopping over OAuth 2.1, plus a generator that compiles any OpenAPI spec into a deployable MCP server with Docker and Kubernetes artifacts
    • Traced every agent step and tool call with OpenTelemetry and Langfuse, and attached content guardrails per agent with violations logged, so a bad run can be traced back to the exact turn that caused it
    • Designed and rolled out a federated OAuth 2.0 and OIDC layer with PKCE, token introspection, and revocation across four backend services
    • Own delivery end to end: architecture planning with senior and principal architects, CI/CD, AWS infrastructure, and deployment. Unblock other teams when they're stuck and run internal engineering sessions
    • Won 2 INSTA Awards in back-to-back quarters for this work
  • I

    Infosys Equinox - Enterprise Digital Commerce Platform

    Systems Engineer

    • Resolved 4,000+ security vulnerabilities across every Equinox microservice in 6 months using Snyk and policy automation, clearing the platform's compliance benchmarks
    • Led the migration of 50+ Spring Boot microservices (2.x to 3.x) and Apache Camel routes (2.x to 4.x) onto Jakarta EE, improving performance by ~25%
    • Automated that migration with a Python tool that rewrote Camel routes and javax-to-jakarta references across 500+ routes, cutting migration time by ~55%
    • Built a Python and Streamlit DevSecOps dashboard for vulnerability triage and per-release security bulletins, cutting average resolution time by ~40%
    • Led a team of 5 developers in an agile POD, running sprint planning and cross-team coordination, and delivered SDK-generation milestones two months ahead of schedule
    • Earned the "Above and Beyond" and "Spotlight Engineer" awards for delivery under pressure

achievements

Infosys ESOP Stock Award

Recognized personally by CEO Salil Parekh for exemplary contribution to AI-first direction (March 2026).

Google DeepMind Hackathon Winner

NanoStudio - one of the winners out of 800+ worldwide submissions at the Nano Banana Hackathon.

NeurIPS 2025 Bronze Medal

78th rank out of thousands at the Google Code Golf Championship.

Spotlight Engineer & Above and Beyond Awards

Infosys Equinox internal recognition for exceptional delivery.

what leaders say

  • “A highly skilled and technically sound professional with exceptional expertise in AI, Agentic AI, and the latest advancements in the AI ecosystem. Mrutyunjay demonstrates outstanding programming, analytical, and problem-solving abilities, consistently delivering innovative and effective solutions. He possesses deep technical knowledge combined with a proactive approach to learning and applying emerging technologies.”

    Pankaj Parandkar

    Senior Technology Architect · Infosys Limited

  • “I had the opportunity to work with Mrutyunjay as his Product Owner on an Agentic AI implementation for the Infosys Equinox Platform, and it was one of the most impressive collaborations I've experienced. Mrutyunjay was the sole technical lead on this initiative, taking complete ownership of both the solution architecture and end-to-end delivery. What set him apart was a rare combination of deep technical expertise, bold decision-making, and the ability to drive complex work entirely independently. A moment that truly defined his contribution was when he identified a fundamental flaw in the existing Agentic Studio architecture mid-flight. Without hesitation, he pivoted to a new architecture and delivered a fully functional AI Studio within a week, a turnaround that earned recognition from stakeholders and peers across the organisation. Beyond the technical brilliance, Mrutyunjay communicated complex AI concepts with clarity, bridged the gap between technology and business outcomes effortlessly, and never needed to be chased or directed. He simply got things done, and got them done exceptionally well. I would wholeheartedly recommend Mrutyunjay to anyone looking for a technology leader who combines architectural depth with the ownership mindset of a founder.”

    Mayank Jain

    Senior Product Manager · Infosys Limited

featured projects

view more
Psych Runtime

Psych Runtime

An open-source, embeddable Python agent runtime under Apache-2.0 - the agent loop, durable execution, MCP with tenant isolation, and a run report, so a company can build its own agentic platform without becoming one.

Python 3.12+
Pydantic v2
MCP
PostgreSQL
OpenTelemetry
AgentCart

AgentCart

An agentic e-commerce platform for clothing and accessories - Maya, an AI shopping assistant, searches, checks out, and generates virtual try-ons through a 7-tool MCP server with semantic search over Cloudflare Vectorize.

Next.js 16
Cloudflare Workers
MCP
Cloudflare Vectorize
D1
NanoStudio

NanoStudio

One of the winners of the Google DeepMind Nano Banana Hackathon, out of 800+ worldwide submissions - turns one product photo into campaign-ready mockups and edits using Gemini 2.5 Flash Image.

Next.js 15
Gemini 2.5 Flash
TypeScript
Turso/LibSQL
Vercel
Equinox Agentic Platform

Equinox Agentic Platform

A standalone, multi-tenant agentic AI platform I built end to end and solo - a FastAPI graph-based agent runtime, orchestration, and the full Next.js studio UI that drives all of it.

FastAPI
LangGraph
LiteLLM
Next.js
MongoDB
Redis

recent blogs

view more