Vraj Shah

AI engineer and full-stack developer

I build LLM systems that hold up in production: agents, evaluation pipelines, and the full-stack products around them.

AI engineer and full-stack developer in Berlin

At 50Hertz I own four internal production systems end to end, from LLM agents for grid operations analysis to the FastAPI and React apps around them. I proposed the evaluation pipeline that now catches regressions in agent output on every CI run.

Before that I founded Chariot Technologies and shipped IoT automation for paying clients: sensors, microcontrollers, cloud services and remote device management.

Now
50Hertz Transmission GmbH Sep 2025 – present
Founded
Chariot Technologies Dec 2022 – Dec 2024

Selected work

Experience & education

  1. Software Developer / Founder

    Chariot Technologies,

    • Built and deployed IoT automation systems: sensors, microcontrollers, cloud services, remote device management.
    • Led architecture, deployment, and client implementation; grew it into a revenue-generating business with ₹1 lakh government funding.

  2. BSc (Hons) Computing

    Arden University, Berlin,

  3. Working Student Software Developer (Full-Stack / AI)

    50Hertz Transmission GmbH (AI Incubator),

    • Designed and implemented an LLM evaluation pipeline (Azure AI Evaluation SDK) running in CI to catch regressions in agent output quality; proposed the approach and had it adopted. Built production LLM agents on Azure AI Foundry: prompt design, tool/function calling, structured output with schema validation, retrieval over internal documents, and prompt-injection guardrails. Own four internal production systems end to end: an LLM agent for grid operations analysis, a forecasting tool feeding regulatory reporting, a tender-analysis application, and a documentation system. Built FastAPI/PostgreSQL backends with React/TypeScript frontends; validation pipelines, filtering, and database integration.

Skills

The stack from the model down to the hardware. The number beside each skill counts the projects and roles that use it; hover, focus or tap a skill to see which.

Models and agents

Getting LLMs to do dependable work.

Services and data

The backends that agents call and products run on.

Interfaces

Where people meet the system.

Delivery and hardware

Shipping it and keeping it running.

Building something with LLMs? Tell me what you are working on.

vraj0410shah@gmail.com

Or send a message here and it lands in my inbox.