AI Engineer • ML + AI Enthusiast • Builder

I like shipping practical ML/GenAI systems end-to-end — from data + modeling to lightweight apps/demos — and turning messy real-world signals into decision-ready insights.

• Applied ML & GenAI: text classification, evaluation loops, prompt/agent workflows, and turning experiments into reproducible pipelines
• Analytics that drives action: EDA → feature engineering → modeling → clear business takeaways
• Product-minded engineering: prototypes that are demo-able, documented, and easy to iterate on

• text_classification
  Text classification experimentation repo with structured outputs (e.g., gpt_results/, hermes_results/, results/) and evaluation notebooks for comparing runs. oai_citation:0‡GitHub

• airbnb-superhost-prediction
  End-to-end Superhost prediction pipeline: preprocessing → feature selection → model development (LogReg, RF, XGBoost, NN) → evaluation (precision/recall, ROC) + key driver analysis. oai_citation:1‡GitHub

• InstagramClone_withFastAPI
  “Simple Social”: FastAPI backend + Streamlit frontend with JWT auth (via FastAPI Users), async SQLAlchemy + SQLite persistence, and ImageKit-backed media uploads (images/short videos) with a unified feed. oai_citation:2‡GitHub

• RAG-Project
  Retrieval-Augmented Generation pipeline over local text docs: load .txt → chunk (1000 chars, 20 overlap) → embed (text-embedding-3-small) → store/retrieve with ChromaDB → cosine-similarity ranked retrieval (top-k) for context-fed QA with GPT models. oai_citation:3‡GitHub

Languages: Python, SQL, (some) Kotlin / Dart
ML / DS: Pandas, NumPy, scikit-learn, XGBoost/CatBoost, Jupyter
GenAI: LangChain-style orchestration, evaluation-first experimentation
Apps / Prototypes: Flask (lightweight demos), notebooks, small tools
Workflow mindset: clean repos, repeatable runs, clear READMEs, measurable results

• Portfolio / Resume Website: https://vishnuanand77.github.io/resume-website/
• GitHub: https://github.com/Vishnuanand77