Daniel Yeke's Work | ContraWork by Daniel Yeke
Daniel Yeke

Daniel Yeke

Python developer building AI agents and SaaS solutions.

    2
    Followers
Followed by Johnson A and StackShortlist

Naaldwijk, Netherlands

WorkPostsAbout

Naaldwijk, Netherlands

Cover image for Autonomous MLOps & Software Engineering
Autonomous MLOps & Software Engineering Agent I built an AI-powered developer agent designed to inspect a codebase, work with project files, execute development commands, and retrieve relevant context from vector memory. Key capabilities Tool-calling agent: Registered tools for listing project files, reading and writing files, and running terminal commands. API backend: FastAPI service for the agent's application interface. Vector memory: ChromaDB integration for semantic retrieval of codebase context and documentation. Interactive dashboard: Streamlit interface for interacting with the agent. Error-correction workflow: Designed to use execution errors as feedback for subsequent correction attempts. Technology stack: Python, FastAPI, Groq, ChromaDB, Streamlit. Project focus: Agentic AI, developer automation, retrieval-augmented generation, and AI application engineering.
0
8
Cover image for Autonomous Vision-Language Navigation Rover Overview Developed a
Autonomous Vision-Language Navigation Rover Overview Developed a simulated autonomous rover combining robotics, computer vision, and sensor-fusion concepts to explore navigation in a simulated environment. Project focus ROS 2 and Gazebo simulation. Camera-based object perception using an ONNX model. Integration of robotics sensors and localization concepts. Exploration of vision-language navigation workflows. Technologies: ROS 2 Humble, Gazebo, Python, ONNX Runtime, computer vision, sensor fusion.
0
13
Cover image for CRAG — Self-Correcting Research Assistant Overview A
CRAG — Self-Correcting Research Assistant Overview A research-assistant project exploring how AI applications can retrieve information and improve the quality of their responses through a self-correction workflow. Project focus Python-based AI application development. Retrieval and research workflow design. Improving answer reliability through evaluation and correction. Technologies: Python, AI, retrieval-augmented generation.
3
1
47
Cover image for Multi-Provider AI Router — Docker
Multi-Provider AI Router — Docker & CI Automation I built a Python API that routes requests across configured AI providers, records routing diagnostics, and falls back to a local mock response when remote providers are unavailable. What I implemented Containerized the FastAPI application with Docker. Added health-check verification. Implemented provider-routing diagnostics and mock fallback handling. Added automated tests for routing behavior, circuit-breaker recovery, and API errors. Configured GitHub Actions to run tests and application checks automatically. Validation Docker image built and the container started successfully. The /health endpoint returned online. The mock fallback returned a response with routing diagnostics. Eight automated tests passed locally, and GitHub Actions CI passed. Scope and limitations: The mock fallback is not generated model output. Groq was tested previously; live Cerebras connectivity has not been verified. Streaming and per-request model selection are not implemented. This project demonstrates my practical work with Python APIs, Docker, automated testing, and CI workflows.
0
28