Skip to content
View niravpatidar37's full-sized avatar
💭
Great Results Require Greater Ambition 🚀
💭
Great Results Require Greater Ambition 🚀
5 followers · 12 following

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx2Achievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx2Achievement: YOLOAchievement: Quickdraw

Block or report niravpatidar37

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
niravpatidar37/README.md

Nirav Patel

AI Engineer · Agentic AI Specialist · Open Source Builder · MCP Security

Typing SVG

LinkedIn Portfolio Email

Profile Views Followers Stars

🚀 "I build GenAI systems that are measurable, debuggable, and resilient under real traffic."

👋 Who I Am

I'm Nirav Patel, an AI Engineer in Toronto with 5 years across software, analytics, and AI. I'm currently an Agentic AI Specialist at Adamson Systems Engineering.

I build GenAI applications end to end with Python, LangGraph, Amazon Bedrock, Claude, and Azure OpenAI, taking them from a customer's problem to production. Before Adamson, I shipped a four-agent production suite and hybrid-search RAG (BM25 + dense vectors + Cohere re-ranking) on AWS Bedrock at DeepIDV.

In open source I work on AI security and code review: mcpsum (a lockfile and runtime monitor for MCP tools, in Rust) and MergeCode (a judge that reviews AI-generated patches the way a maintainer would).

⚡ At a Glance

name:        Nirav Patel
role:        Agentic AI Specialist
company:     Adamson Systems Engineering
location:    Toronto, ON
focus:       Agents · RAG · MCP security
experience:  5 yrs in software & AI
languages:   Python · Rust · TypeScript
clouds:      AWS Bedrock · Vertex · Azure
oss:         mcpsum · MergeCode · VibeForge
principle:   Clean abstractions > demos

🔭 Currently Building & Learning

🛠️ Building

  • 🔐 mcpsum: lockfile + runtime reference monitor for MCP tools (Rust)
  • ⚖️ MergeCode: maintainer-grade review for AI-generated code
  • 🕸️ Graph RAG: Qdrant + Neo4j retrieval with evals and Langfuse tracing
  • 📒 A trading journal with a multi-agent research desk

🌱 Learning & Exploring

  • 🦀 Rust for security-critical agent tooling
  • 🛡️ MCP & agent security: tool poisoning, rug pulls, prompt injection
  • 🧪 LLM evals and red-teaming in CI
  • 🏠 Local-first LLMs with Ollama on consumer GPUs

🧰 Tech Stack

python    aws    docker    react

🧠 Languages & Core python rust typescript nodejs bash git linux
🤖 AI · LLMs pytorch hugging face langchain claude ollama
☁️ Cloud · Deploy aws gcp azure docker github actions fastapi streamlit
🗄️ Data · Retrieval postgres redis neo4j qdrant
🧩 Apps · Automation react nestjs n8n

🧬 GenAI · LLMs · Agents · Vector DBs

LangGraph LangChain Claude Amazon Bedrock Azure OpenAI Vertex AI Hugging Face Ollama MCP Qdrant Neo4j Cohere Rerank Langfuse

I build agents by day and evaluate them by night 🤖📈

📌 Pinned Projects

Project What it does Stack
mcpsum Stops MCP servers from rug-pulling your AI agent: a lockfile + runtime reference monitor for MCP tools (like go.sum, for the tools your agent trusts) Rust · MCP
VibeForge Mood-to-playlist agent: Mood Analyst → Music Curator → self-correcting Critic loop (featured in 500 AI Agents Projects) LangGraph · Groq · FastAPI · Redis
MergeCode Maintainer-grade review for AI-generated code: runs your checks, catches the tricks that get patches to green (deleted asserts, skipped tests, new deps, CI edits), returns MERGE / REQUEST_CHANGES / REJECT with evidence TypeScript · Node.js
AI-Task-Architect Turns natural-language prompts into importable n8n workflow JSON NestJS · FastAPI · GPT-4o · n8n

📊 AI Usage Embed

Tokscale Stats

📊 GitHub Metrics & Stats Dashboard

github stats top langs streak stats

📈 Contribution Graph

Activity Graph

🐍 Contribution Snake

Snake animation

📫 Connect With Me

GitHub LinkedIn Email

🎯 Current Focus

🤖 Day job       AI Engineer @ Adamson
🔭 Side project  mcpsum — lockfile + runtime monitor for MCP tools (Rust)
📒 Building      My own trading journal with a multi-agent research desk
💬 Ask me about  MCP security, LangGraph, RAG, Multi-Agent Systems, Amazon Bedrock
⚡ Fun fact      I build agents whose job is to critique other agents
🧭 Principle     Clean abstractions > fancy demos

💡 Random Dev Quote

Readme Quotes

Retrieve. Reason. Reflect. Repeat.

🔁 The graph always has one more node.

Pinned Loading

  1. mcpsum mcpsum Public

    Stop MCP servers from rug-pulling your AI agent. A lockfile + runtime reference monitor for MCP tools.

    Rust 2 1

  2. vibeforge vibeforge Public

    VibeForge — AI agent that forges a 10-track playlist from your mood. LangGraph critic loop, FastAPI + React, Groq / Hugging Face.

    Python 1

  3. mergecode mergecode Public

    Maintainer-grade review for AI-generated code. Runs your checks, catches the tricks that get patches to green (deleted asserts, skipped tests, new deps, CI edits), and returns MERGE / REQUEST_CHANG…

    TypeScript

  4. AI-Task-Architect AI-Task-Architect Public template

    Generates structured n8n workflow JSON from text prompts using GPT-4o, orchestrated via a NestJS Gateway and FastAPI AI Engine.

    Python