Skip to content
View teotia-codes's full-sized avatar
🎯
Focusing
🎯
Focusing
9 followers · 6 following

Achievements

Achievement: YOLOAchievement: Pull Sharkx2

Achievements

Achievement: YOLOAchievement: Pull Sharkx2

Block or report teotia-codes

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 is supported. This note will only be visible to you.
Report abuse

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

Report abuse
teotia-codes/README.md

Typing SVG


Profile Views   LinkedIn   GitHub

$ whoami

class PriyanshuTeotia:
    def __init__(self):
        self.name        = "Priyanshu Teotia"
        self.degree      = "Integrated M.Tech CSE (Business Analytics) — VIT Chennai"
        self.cgpa        = 8.95
        self.role        = "AI Engineer | LLM Infra | Agentic Systems | India 🇮🇳"

        self.building    = [
            "AutoInsight AI   →  FastAPI + LangGraph + ChromaDB + Ollama + React",
            "RAG Pipelines    →  HuggingFace Embeddings + ChromaDB + Semantic Retrieval",
            "Fine-tuned LLMs  →  TinyLlama 1.1B + LoRA Adapter for Report Generation",
        ]

        self.core_stack  = [
            "Python", "FastAPI", "LangGraph", "LangChain",
            "OpenAI-compatible APIs (Ollama/Qwen)", "ChromaDB", "HuggingFace",
            "PostgreSQL", "MongoDB", "asyncio", "Docker",
        ]

        self.ask_me      = ["RAG Pipelines", "Agentic AI", "LLM Fine-Tuning", "FastAPI Async"]
        self.looking_for = "AI Engineer Internship — open to Bangalore / Hybrid"
        self.motto       = "Ship fast. Think deep. Build things that matter."

    def __repr__(self):
        return "Turning raw data into deployable AI intelligence."

$ cat tech_stack.json

🧠 AI · LLM · Agentic Engineering

Python PyTorch HuggingFace LangChain LangGraph Ollama

⚙️ Backend · APIs · Async

FastAPI asyncio REST APIs Docker Streamlit

🗄️ Vector DBs · RAG · Databases

ChromaDB PostgreSQL MongoDB MySQL Redis

📊 ML · Data Science

scikit-learn Pandas NumPy Power BI SHAP

$ ls featured_projects/

🤖 AutoInsight AI

End-to-end agentic data intelligence platform

  • 🔗 10-node LangGraph multi-agent pipeline — profiling, KPI extraction, data quality auditing, ML readiness
  • 🧠 Full RAG pipeline — chunking, HuggingFace embeddings, ChromaDB vector indexing, semantic retrieval
  • 🦙 Fine-tuned TinyLlama 1.1B + LoRA adapter wired into report generation node with Ollama fallback
  • ⚡ Async FastAPI backend with in-memory caching to minimize redundant LLM calls
  • 🖥️ React frontend with SSE-based progress streaming + multi-format file upload (CSV, TSV, Excel, JSON)

Stack: FastAPI · LangGraph · ChromaDB · Ollama · HuggingFace · React

🩺 Explainable AI Diabetes Risk Predictor

Full-stack ML inference system with interpretability

  • 📈 End-to-end time-series ML pipeline — feature engineering, model training, validation (Scikit-learn)
  • 🚀 FastAPI RESTful inference service with MongoDB data store and scalable model serving
  • 🔍 SHAP-based Explainable AI for interpretable clinical prediction dashboards
  • 📱 Cross-platform Flutter app consuming backend prediction APIs
  • 🐳 Docker + GitHub Actions CI/CD for automated testing and deployment

Stack: FastAPI · Scikit-learn · SHAP · MongoDB · Flutter · Docker

📈 Clinical Risk Analytics Platform

22.6M+ record clinical glucose risk pipeline

  • 🗄️ MySQL → Python → Power BI end-to-end analytics pipeline
  • 🏗️ SQL-based feature engineering with window aggregations (mean, max, variability, monitoring intensity)
  • 🤖 Logistic Regression + Random Forest achieving ROC-AUC 0.87, Accuracy 82%
  • 📊 Executive Power BI dashboard with star schema, dynamic KPIs, and drill-through reporting

Stack: Python · MySQL · Scikit-learn · Power BI · SQL

🎥 Mobile Deepfake Detection — Edge AI

3D CNN deepfake detection optimized for on-device inference

  • 🎯 96% accuracy, 75% recall, 432ms inference on resource-constrained mobile devices
  • 🖼️ Computer vision techniques: CLAHE, ELA, optical flow, noise residual analysis
  • ⚡ Optimized for real-time edge inference — no cloud dependency
  • 🏗️ Object-oriented ML architecture with version-controlled workflows

Stack: PyTorch · Python · 3D CNN · Edge AI · Git

$ cat achievements.log

[2024] 🏆 Smart India Hackathon (SIH) — TOP 5 FINALIST (National Level)
       Built data-driven agricultural marketplace for farmer-retailer direct trade.
       Secured Top 5 after multiple national evaluation rounds.

[2024] 💼 Data Analyst Intern — Jivan Health (Health-tech AI Chatbot Platform)
       → Python + SQL EDA on healthcare interaction datasets
       → Power BI KPI dashboards: patient engagement, retention, response latency
       → Automated reporting pipelines, reduced manual reporting effort

$ git log --stat



 

$ watch contribution-graph

contribution snake

$ trophy --display-all

$ tail -f currently_learning.log

> Exploring RLHF concepts and reward modelling for LLM alignment
> Deepening knowledge of Kafka / Celery for async AI pipeline orchestration  
> Experimenting with Google ADK for multi-agent workflows
> Advancing GPU training optimization and CUDA-level ML engineering

"The best AI engineer is the one who ships."

Popular repositories Loading

  1. FarmlyApp FarmlyApp Public

    Dart 2 1

  2. adv_basics adv_basics Public

    C++

  3. first_app first_app Public

    C++

  4. expense_app expense_app Public

    C++

  5. calculator calculator Public

    C++

  6. meals meals Public

    Meals app showcasing the steps required to prepare a meal , all indicating all the indegrients required in preparation of the selected meal. Riverpod package is used for cross widget management.

    Dart