Hi, I'm Keval Savaliya 👋

AI & Software Engineer · MSc Automotive Software Engineering · TU Chemnitz, Germany

---

🧠 What I Built

I build production-ready AI systems — from RAG pipelines and multi-agent architectures to MCP servers and ML REST APIs. My focus is on AI that is reliable, integrable, and actually ships.

focus = {
    "current":     "MCP server development for AI workflow integration",
    "specialties": ["LLMs", "RAG", "Multi-Agent Systems", "REST APIs", "CI/CD"],
    "languages":   ["Python", "TypeScript", "C#", "SQL"],
    "domain":      "Clinical Software · Safety-Critical Systems · AI Tooling"
}

---

🚀 Featured Projects

🏥 MCP Clinical Workflow Assistant

An MCP server exposing AI tools for clinical software development — FHIR resource generation, medical abbreviation resolution, and IEC 62304 compliance checklists. Connects directly to Claude Desktop.

Python MCP FHIR Claude Desktop Pydantic

---

🤖 MCP Dev Workflow Tools

An MCP server that plugs into development workflows — automated code review checklists, conventional commit message generation, and API documentation formatting.

Python MCP Claude Desktop FastAPI

---

🔍 RAG + Multi-Agent Knowledge System

AI-powered assistant using Retrieval-Augmented Generation with a modular multi-agent architecture. Improved response reliability by 40% over a single-agent baseline.

Python LangChain Vector DB LLMs Prompt Engineering

---

🧬 BERT NLP Classification System

Text classification system using fine-tuned BERT transformers, achieving 81.5% F1-score. Includes hyperparameter tuning, evaluation pipeline, and a Streamlit demo UI.

Python PyTorch Transformers Streamlit scikit-learn

---

⚙️ End-to-End ML Service with REST API

Production-ready ML pipeline exposed via FastAPI, containerized with Docker. Real-time predictions with a clean architecture designed for extensibility.

Python XGBoost FastAPI Docker REST API

---

🛠️ Tech Stack

Area	Technologies
AI / ML	PyTorch, TensorFlow, Transformers, LangChain, RAG, LLMs, MCP
Backend	FastAPI, Flask, REST APIs, Microservices
DevOps	Docker, Kubernetes, Jenkins, CI/CD, GitHub Actions
Data	SQL, ETL Pipelines, Data Validation
Languages	Python · TypeScript · C# · Java · SQL

---

📈 Experience

• Data Scientist Intern @ Data Glacier (2025 – Present) — Python automation tools, REST APIs, 60% reduction in manual processing
• Data Scientist Intern @ HackSec Infotech (2023 – 2024) — CI/CD pipelines, Docker/Kubernetes, 30% faster deployments

---

---

"Building AI systems that work in the real world — reliable, interoperable, and production-ready."