Hi👋, I'm Abylay. Here is the a structured view on things I did or am doing or tried to do or trying to do. Some deployed, some on github.
Langchain, Gemini API, Telegram Bot API, Mistral OCR
RAG system for answering user's legal questions by searching through Codes of Law of Kazakhstan.
Langgraph
In Progress... Agentic system to analyze professional DOTA 2 matches (hero picks, teams).
HuggingFace, trl, lm-eval, detoxify
Fine-tuning open-source LLM dolly-v2-3b model with DPO algorithm with Runpod cloud A100 80GB GPU
Python, Selenium, Sklearn, XGBoost, CatBoost, LightGBM, PostgreSQL, SHAP, Pandas, Matplotlib, Seaborn, FastAPI
End-to-end ML project to predict the outcomes of professional CS2 matches.
Python, OpenAI API, ChatGPT4o-mini, Docker, LlangChain, Telegram Chatbot
A project I made for fun with OpenAI API to query the web with gpt4o-mini.
Python, Selenium, Sklearn, XGBoost, CatBoost, LightGBM, PostgreSQL
Collected a new dataset - Same task.
Python, Tensorflow, Keras
Second place at Hackathon. Identifying fire on images.
Pytorch, HuggingFace Transformers, HuggingFace Datasets, HuggingFace Metrics
Solving various Computer Vision tasks. Mainly for practice or testing some ideas and being up-to-date.
Pytorch, Tensorflow, NLTK, HuggingFace Transformers
Solving various NLP tasks.
Tensorflow, Keras
Some exercises I am doing while practicing Tensorflow/Keras
Python, Scipy, Statsmodels
Some Statistics with Python.
Python, Matplotlib, Seaborn
Some Data Visualizations with Python.
Python, Sklearn, XGBoost, CatBoost, LightGBM, Optuna, Matplotlib, Seaborn
Got top5% submission. Focused on ensemble methods: stacking, blending, voting, ranking