👋 Hi, I'm Niteesh ( @niteesh_ai )

Summer of AI 2025 – Viswam.ai | Swecha
Building Indic datasets and fine-tuning LLMs.

🚀 About Me

• AIML student (3rd year), India
• Interests: NLP, Flutter, Data Engineering, MLOps
• Currently: Corpus collection & LLM fine-tuning for Summer of AI 2025

🧰 Skills

Languages: Python, Java, C/C++
AI/DS: NumPy, Pandas, Scikit-learn, PyTorch/TensorFlow (basic)
Dev: Git, GitLab, Linux CLI, Docker (basic)
Mobile: Flutter/Dart (beginner–intermediate)

🎯 Internship Goals (2025)

• 80+ hours Audio/Video corpus
• 800+ Image/Text records
• Build a Collaborative Corpus Collection Engine (team project)
• Fine-tune an LLM for an Indic use-case
• Weekly updates in this README

🗂️ Quick Links

• Code: code.swecha.org/<your-username>
• Chat (Mattermost): chat.swecha.org
• Courses: courses.viswam.ai
  • Python Simple | Python Advanced | Learn AI
• Corpus App/Web: corpus.swecha.org (and Android app)

📚 Learning Path (Live Checklist)

Week 1–2

• Python Simple (basics, functions, files)
• Git/GitLab workflow, Issues/MRs
• Start corpus: 20h A/V + 200 Image/Text records

Week 3–4

• Python Advanced (OOP, generators, venv, packaging)
• Data cleaning + labeling pipeline
• Another 20h + 200 recs

Week 5–8

• Learn AI (NLP basics, tokenization, embeddings)
• Fine-tune small LLM on collected corpus
• Deployment notes, evaluation metrics

📦 Projects

• Corpus Collection Engine (Team):
  Role: Data pipeline & validation
  Stack: Python, FastAPI, Postgres, MinIO, Docker
  Status: 👇 (update weekly)

• LLM Fine-tuning (Indic):
  Model: 👈 e.g., Llama-3-Instruct
  Task: 👈 e.g., Q&A in Telugu
  Data: My collected corpus (cleaned + split)
  Metrics: 👈 e.g., BLEU/ROUGE/Accuracy

📊 Corpus Tracker (update daily)

Totals

• Audio/Video: 0 / 80h
• Images/Text: 0 / 800

📝 Daily Log

• 2025-08-23: Set up README, enrolled in Python Simple, 2h audio collected.
• 2025-08-24: 👈

🧪 Tech Notes (snippets I reuse)

• Virtual env: python -m venv .venv && source .venv/bin/activate
• Lint/format: ruff . && black .
• Train script sample: python train.py --epochs 3 --lr 3e-5 --batch 8

📫 Contact

• Email: [email protected]
• Chat handle (Mattermost): @👈
• LinkedIn/GitLab: @niteesh_ai

This README is my living portfolio for Summer of AI 2025. I’ll update it weekly with progress and artifacts.