Machine Learning Engineer
CurrentMachine learning engineering, data analysis, ad-hoc data research:- Developed and implemented containerized API-based LLM RAG assistant system (fastapi, langchain, faiss etc.)- Developed and implemented classification model (lightGBM) to predict enrollment (from idea to MVP) with F1 = 69%- Researched and developed the demand rating for digital education materials in the electronic library- Implemented and fine-tuned LLMs (BERT, roBERTa) to classify digital education materials (from idea to MVP streamlit + fastapi app) with weighted F1 ~ 0.6- Organized the work of Big Data and Machine Learning department from scratch (together withsuperior)- Organized ETL processes for massive amount of data from 8 different systems; collected, cleaned and transformed this data to assemble ML datasets- Provided ad-hoc data analysis and vizualization for various stakeholdersStack:- Linux, jupyter lab, git, docker- Python: pandas, numpy, sklearn, langchain, huggingface, pytorch, seaborn, matplotlib, bs4, statsmodels, lightgbm, xgboost, scipy, FastApi, dask, sqlalchemy etc.- ML: classic ML, deep learning NLP (LLMs)- Data: SQL (window functions, query optimizing), web scrapping, API, Apache Drill, vector stores