Skip to content
View KonNik88's full-sized avatar
5 followers · 1 following

Block or report KonNik88

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
KonNik88/README.md

Hi there

I’m Konstantin Nikiforov — MD, Molecular Geneticist (Russia) transitioning into Data Science & Machine Learning.
I like end-to-end work: data → features → models → APIs → simple UIs → orchestration.

Focus: applied ML for business (analytics & decision support), reproducible pipelines, model calibration & evaluation.
Interests: neural networks & AI, biomedical ML, causal inference, interpretable models, CV, NLP/linguistics, music, Time Series.

Stack

Core: Python · SQL · Pandas · NumPy · scikit-learn · CatBoost · XGBoost · LightGBM · Optuna
DL/NLP: PyTorch · Hugging Face · Sentence-BERT · (SimCLR/BYOL, Diffusion — exploring)
Recsys: ALS (implicit) · Hybrid SBERT+ALS+CatBoost · Qdrant (vector DB)
Time Series: Prophet · TBATS · ETNA/AutoTS · backtesting (rolling/holdout)
MLOps/App: FastAPI · Streamlit · Airflow · Docker/Compose · MLflow
Data & Scale: Spark (PySpark) · PostgreSQL · MySQL
XAI/Visualization: SHAP · LIME · Plotly/Dash · Matplotlib · ydata-profiling
Domains: Tabular ML · Recommenders · NLP · Time Series · BioML · CV

Exploring: Ray/Dask · GNNs · SSL · ESM · AudioML · Rust · C/C++ · Java

Selected Projects

  • Hybrid Book Recommender System — CatBoost + ALS + SBERT · FastAPI + Streamlit · Docker · Qdrant
    repo
  • BlendCAL — Conversion Prediction — CatBoost/XGBoost/LightGBM ensemble · FastAPI · Streamlit · Airflow DAGs · Docker Compose
    repo
  • Model Drift Monitoring — Evidently + SHAP + PSI/JS checks · alert policy demo
    repo
  • Panel Time-Series Forecasting — ARIMA, TBATS, Prophet, Darts · Optuna-tuned baselines
    repo
  • Omics Survival Analysis — RNA-seq PCA + embeddings for bioinformatics
    repo

Languages

English (B2) · French (B2)

Contact

Email: konnik1000@gmail.com · Telegram: @Konnik1988 · GitHub: https://github.com/KonNik88

TL;DR

MD molecular geneticist building practical DS/ML pipelines (Python, FastAPI, Streamlit, Airflow, Docker).

Pinned Loading

  1. omics-survival-embeddings omics-survival-embeddings Public

    Benchmarking embedding methods (UMAP, VAE, PCA, FA, ICA, etc.) for survival prediction on omics data with TabNet, CatBoost and ridge models.

    Jupyter Notebook 4

  2. audio-similarity-tagging-hub audio-similarity-tagging-hub Public

    Universal audio embeddings + tagging + similarity search with Streamlit demo and FastAPI; PANNs/YAMNet, FAISS/Qdrant, Grad-CAM.

    Jupyter Notebook 1

  3. hero-portrait-generator hero-portrait-generator Public

    Hero Portrait Generator — GAN/VAE/DDPM project A creative deep learning project for generating fantasy-style hero portraits (inspired by retro pixel-art / RPG games)

    Jupyter Notebook

  4. ml-research-radar ml-research-radar Public

    Custom end-to-end pipeline to find, organize, and reason over ML papers & GitHub repos — with an experimental NVIDIA UDR compare tab

    Python

  5. multimodal-retrieval-lab multimodal-retrieval-lab Public

    Production-ready multimodal retrieval system built with OpenCLIP, Qdrant, FastAPI and Streamlit. Includes full evaluation pipeline (Recall@K, mAP, nDCG) and Docker-based deployment.

    Jupyter Notebook

  6. ssl-cifar10-simclr-byol ssl-cifar10-simclr-byol Public

    Self-Supervised Learning on CIFAR-10 with SimCLR and BYOL. Reproducible PyTorch pipeline with linear probe, k-NN evaluation, and visualization

    Jupyter Notebook