Learned #RAG and #Agents for using the #LLM Api with free #LLMZoomCamp course of #DataTalksClub .

๐Ÿš€ Just finished my #DEZoomcamp project! I built an end-to-end pipeline to process population frequencies from the European Variation Archive (EVA).

The Stack:
๐Ÿ› ๏ธ Orchestration: #Bruin (Asset-based & lightweight)
๐Ÿ“ฅ Ingestion: On-the-fly Python filtering
โšก DWH: #ManticoreSearch for sub-second variant lookups
๐Ÿ“Š UI: #Gradio dashboard
๐Ÿณ Env: #Docker & Codespaces & Cloud Run

Efficiency > Big Budgets. ๐Ÿงฌ

๐Ÿ”— https://github.com/tnotstar/data-engineering-zoomcamp-2026-project-attempt-1

#DataEngineering #Python #OpenSource #LearningInPublic #DataTalksClub

GitHub - tnotstar/data-engineering-zoomcamp-2026-project-attempt-1: First project attempt for Data Engineering Zoomcamp

First project attempt for Data Engineering Zoomcamp - tnotstar/data-engineering-zoomcamp-2026-project-attempt-1

GitHub
๐Ÿงช Phase 7: Integration Testing
The final touch: integration testing for all pipelines! Ensuring smooth functionality between training, deployment, and monitoring ๐Ÿšฆ. #MLOpsZoomcamp #DataTalksClub
๐Ÿ“Š Phase 6: Model Monitoring
Set up the monitoring pipeline for detecting data and model drift with Evidently ๐Ÿ”Ž. Grafana dashboards are live, and Iโ€™m tracking model performance in real time! #MLOpsZoomcamp #DataTalksClub
๐Ÿš€ Phase 5: Deploying the Model
Deploying the model using BentoML ๐Ÿง‘โ€๐Ÿ’ป. Serving it as a scalable API in Docker for production. Everything is automated and ready to go live! #MLOpsZoomcamp #DataTalksClub
โš™๏ธ Phase 4: Model Training Pipeline
Finalizing the training pipeline for my optimized model ๐Ÿ…. Itโ€™s now tracked in MLflow and promoted to production! Letโ€™s prepare it for deployment. #MLOpsZoomcamp #DataTalksClub
๐Ÿ› ๏ธ Phase 3: Tech Stack Setup
All set up with ZenML, MLflow, Optuna, and Docker-Compose for this MLOps project ๐Ÿงฐ. Now the integration begins for seamless pipeline orchestration and experiment tracking! #MLOpsZoomcamp #DataTalksClub
๐Ÿ”ง Phase 2: Model Training & HPO
Feature engineering and model training in full swing! Using XGBRegressor for bike trip predictions ๐Ÿšด and optimizing with Optuna to get the best-performing model. #MLOpsZoomcamp #DataTalksClub
๐ŸŒ Phase 1: Data Exploration
Kicking off my MLOps journey by exploring CitiBike and weather data in NYC ๐ŸŒง๏ธ. Performing some EDA and cleaning the dataset to build the foundation for my prediction model of bike trips. #MLOpsZoomcamp #DataTalksClub
Putting everything to the test by evaluating a dataset using hit rate, MRR, and different approaches like Minsearch and Qdrant. ๐Ÿงช Can't wait to see how they perform! ๐Ÿ“Š #LLMZoomcamp #DataTalksClub