class Soham:
role = "Software Developer | AI Agent & MCP Engineer"
location = "India"
focus = ["Agentic AI", "Model Context Protocol (MCP)", "LLM applications", "Automation"]
languages = ["Python"]
currently = "Building multi-agent workflows and tool-integrated AI systems"
open_to = ["Internships", "Full-time roles", "Open-source collaboration"]I build AI agents that call tools, remember context, and do real work. My projects explore agent orchestration, MCP servers, helpdesk automation, and AI-powered backend services in Python.
| Project | Description | Stack |
|---|---|---|
| gatekeeper | Access-control and validation layer for AI and backend workflows. | Python, Docker, Observability |
| Agentic-Helpdesk | Agent-driven helpdesk that triages and resolves support requests. | Python, LLM agents |
| Agents-Assemble | Multi-agent system where specialised agents collaborate on tasks. | Python, multi-agent |
| neuro_mcp_v2 | MCP server exposing tools for text analysis, memory, and workspace access. | Python, MCP |
Replace each description with a one-line result or metric, for example "cut triage time by 40%". Recruiters and ATS filters both reward measurable outcomes.
| Category | Technologies |
|---|---|
| Languages | Python, SQL, Bash |
| AI / LLM | Agentic workflows, Prompt engineering, Tool/function calling, RAG, Model Context Protocol (MCP), Langgraph |
| Backend | REST APIs, FastAPI, Asynchronous programming |
| Data & Storage | Vector databases, PostgreSQL, SQLite |
| DevOps & Tools | Git, GitHub Actions, Docker |
Keep only what you can discuss in an interview. Edit this table to match your real skills.
- 🦈 Pull Shark: GitHub achievement for merged pull requests
- Add certifications, hackathons, publications, or rankings here as plain text so ATS parsers can read them
Email: so.mukh06@gmail.com · LinkedIn: https://www.linkedin.com/in/soham-mukherjee-787798321