Biotechnology Engineering Graduate · Master's in Biotechnology and Health at ENSB, Algeria.
I build AI systems at the intersection of clinical data, signal processing, and bioinformatics.
- Clinical AI — psychiatric/disease classification from EEG and multimodal clinical data
- Bioinformatics — omics data integration and biomarker discovery pipelines
- Biosignal & biosensor processing — audit, preprocessing, and quality-control pipelines
- Medical device software — DICOM/3D-printing workflows for radiotherapy tooling
- Fine-tuning and deploying LLMs as clinical assistants
| Project | Description |
|---|---|
| BioPrep-AI | LLM-driven biosensor preprocessing tool — audits raw signal/image files, generates and self-tests a custom preprocessing pipeline, containerized with offline-safe fallbacks. Built for OpenAI Build Week 2026. |
| AffiBio | Gene and variant analysis tool combining ClinVar/PubMed literature evidence (Affinage) with an LLM biomedical reasoning agent (Biomni). Analyzes genes, phenotypes, or VCF files via CLI or Streamlit, with dynamic tool selection and structured output enforcement to catch unverified agent claims. |
| NEUROPREDAI | EEG-based psychiatric classification platform (MDD/ADHD/SMC/Healthy) with 85.4% accuracy. Full-stack deployment (Next.js, FastAPI, Supabase) plus a fine-tuned Local clinical LLM assistant. |
| color-app | Streamlit tool to generate complementary, triadic, and analogous color palettes from a name or HEX code, for designers and creatives. |
Python Streamlit Gradio FastAPI Next.js PyTorch OpenCV scipy ML Docker Supabase Hugging Face Ollama DL
- LinkedIn: kharfi-islam-052b94315
- Hugging Face: youslam
- NEUROPRED: neuropredai.com
Open to freelance work and collaborations in clinical AI and bioinformatics.