Available for work · Open to remote and relocation

Satyam Yadav

AI Engineer

Generative AI · Agentic Systems · RAG · MCP

I build production agentic AI: multi-agent systems, RAG pipelines and MCP servers, backed by strong .NET (ASP.NET Core, C#) backend experience.

Bhopal, Madhya Pradesh, India
AI Engineer · SAA Consultancy (UK)

Download Resume Contact GitHub LinkedIn

9merged open-source PRs
MCPofficial C# SDK contributor
5LangGraph agents built
AI-102Azure AI Engineer certified

About

AI Engineer building production multi-agent systems, RAG pipelines, MCP servers and custom computer-vision models with Python, LangGraph, FastAPI, PyTorch and OpenAI. 9 merged open-source PRs, including the official Model Context Protocol C# SDK. Microsoft Certified Azure AI Engineer (AI-102) with a strong .NET backend base.

Based in: Bhopal, Madhya Pradesh, India
Status: Available for work · Open to remote and relocation
Email: [email protected]

Experience

AI Engineer
Dates: Jul 2026 – Present · Remote
SAA Consultancy (UK)
- Enterprise AI
- Agentic Systems
- RAG
- MCP
Agentic Software-Delivery Platform
brief → specification → user stories → tasks
- Built 5 LangGraph agents in Python and FastAPI that turn a project brief into a specification, user stories and typed tasks, with each step proposed by the agent and approved by a person.
- Grounded every extracted requirement in its source text with deterministic quote checks, conflict detection and clarification questions, so unsupported claims are flagged instead of published.
- Built the Control Plane in FastAPI and PostgreSQL with immutable versioning, version diffs, idempotent APIs, row-level tenant isolation, outbox events and human approval gates.
- Implemented a Temporal delivery runtime with human approval stages, and fenced PostgreSQL checkpoints so agent runs recover after a worker crash without duplicate work.
- Recorded token usage for every model call, including failed and timed-out calls, in an append-only ledger with budget checks and an audited admin fix for unknown usage.
- Wrote 650+ automated unit, contract and integration tests, and verified each agent with live runs on the real model.
AI Data Orchestration Platform
agentic ETL workflow automation
- Built the multi-agent ETL automation engine with Python, LangGraph, FastAPI and OpenAI, using planner, executor and reflection agents to generate SQL and Python transformations from relational schemas.
- Designed human-in-the-loop approval and artifact lifecycle so every AI-generated SQL script is reviewed before execution.
- Implemented persistent agent memory in MongoDB (sessions, checkpoint recovery, entity graph) and a Qdrant RAG pipeline grounding agent outputs in schema metadata.
- Developed the platform's MCP servers exposing enterprise APIs and internal tools to LLM clients such as Claude, with an OAuth 2.1 proxy on Azure AD.
Document Intelligence Platform
tender & identity document AI
- Built Identity Document Recognition Service: trained a custom YOLO26 deep-learning model (PyTorch / Ultralytics, Google Colab) that detects and classifies UK passports, driving licences and residence permits, then extracts their text with OCR and passport MRZ parsing.
- Delivered document classification plus Section Planner and Section Writer agents that plan and draft tender document sections.
.NET Full Stack Developer Intern
Dates: Feb 2026 – Jun 2026 · Remote
SAA Consultancy (UK)
- ASP.NET Core
- C#
- SQL Server
- REST APIs
- Built ASP.NET Core REST APIs with C#, SQL Server, Entity Framework Core and Dapper, including JWT auth and role-based access control.
- Integrated FastAPI AI services into existing ASP.NET Core applications through Azure AI Foundry and REST APIs.
- Optimized SQL queries and stored procedures; resolved production issues in an Agile, Git-based workflow.
Full Stack Developer Intern
Dates: Jul 2024 – Jan 2025 · Remote
ThinkBi Tech Solutions Pvt. Ltd.
- .NET Core
- SQL Server
- REST APIs
- Swagger
- Improved page load times by 30% across 10+ .NET Core web applications; built secure REST APIs with Swagger and role-based authentication.

Open Source

9 merged PRs

Official MCP SDKMerged

Model Context Protocol C# SDK

modelcontextprotocol/csharp-sdk

Made McpClient.ServerInfo nullable so .NET MCP clients no longer crash when a server omits optional metadata.
- #1903

Merged

AnythingMCP

HelpCode-ai/anythingmcp

Built the amcp TypeScript CLI with API-key login, named profiles, and MCP tool list/describe/call.
- #1024
- #1088
Added a MOCO REST connector with 8 tools and contract tests.
- #906
Generated llms.txt from repository docs.
- #1199

Merged

NimbleBrain

NimbleBrainInc/nimblebrain

4 fixes to an AI agent engine: rate-limit Retry-After, tool-permission warnings, thinking-tier warnings, Windows build scripts.

In review

View all on GitHub

Skills

Generative AI & Agents

LangGraph
Multi-Agent Orchestration
Agentic AI
RAG
Model Context Protocol (MCP)
Human-in-the-loop Workflows
Agent State & Memory Management
Tool / Function Calling
Prompt Engineering
Vector Embeddings

LLMs & Platforms

OpenAI GPT
Claude
Gemini
Azure AI Foundry
Azure AI Services

AI Backend

Python
FastAPI
Temporal (distributed workflows)
Alembic
REST APIs
C#
ASP.NET Core
TypeScript

Vector & Databases

Qdrant
Pinecone
MongoDB
PostgreSQL
SQL Server

ML, CV & Document AI

PyTorch
YOLO (Ultralytics)
Deep Learning
Computer Vision
Object Detection
OCR
Passport MRZ parsing
NLP
Document Classification

Cloud & Tools

Azure
Git
GitHub
Google Colab
Unit & Integration Testing

Familiar with

CrewAI
AutoGen
MLflow
Databricks
AWS
Neo4j

Education & Certifications

B.Tech, Artificial Intelligence & Machine Learning

Technocrats Institute of Technology & Science, Bhopal

Nov 2022 – Jun 2026 · CGPA 7.23/10

AI-102

Microsoft Certified: Azure AI Engineer Associate

Jun 2025

AZ-900

Microsoft Certified: Azure Fundamentals

Mar 2025

Contact

Hiring for an AI engineering role, or want to talk about agents, RAG or MCP? Send a message or reach me directly.