Skip to content

hawkh/hawkh

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 

Repository files navigation


> whoami

class SaiRuthvik:

    name        = "Sai Ruthvik"
    alias       = "hawkh"
    location    = "IIT Madras, Chennai → Hyderabad, India 🇮🇳"
    email       = "ruthvikworking@gmail.com"

    education   = "B.S. @ IIT Madras  (Final Year)"
    roles       = [
                    "AI/ML Engineer       @ Cyepro Solutions",
                    "Founding ML Engineer @ Visionary EdTech",
                    "Tech Lead            @ Infin AI Club, IIT Madras",
                  ]

    domains     = ["Computer Vision", "NLP & Indic AI", "RAG Systems",
                   "Edge AI", "Audio ML", "ML Systems & MLOps"]

    currently_building = {
        "🌐 Visionary"    : "Multilingual RAG pipeline — 12+ Indic languages, zero LangChain",
        "🚗 Cyepro"       : "CRM-native lead scoring for automobile industry",
        "📐 Side Project" : "Multi-agent math tutoring AI for JEE-level students",
    }

    philosophy  = "Build AI that works for the next billion — not just in English."
    open_to     = ["Founding ML roles", "AI research collabs", "Competitive ML"]

> currently_building

🌐 Multilingual Indic RAG

@ Visionary EdTech

A production RAG pipeline serving 12+ Indic languages with zero framework overhead.

  • No LangChain · LlamaIndex · LangGraph
  • docling → chunking → hnswlib + rank_bm25
  • RRF fusion for hybrid retrieval
  • Ollama local inference on Colab T4
  • GridSearchCV over chunk size & embeddings

🚗 Automobile Lead Scoring

@ Cyepro Solutions

CRM-native ML pipeline to score and prioritize leads for auto dealerships.

  • Feature engineering from CRM event logs
  • XGBoost / LightGBM scoring engine
  • SHAP explainability for sales teams
  • Stage-wise funnel probability outputs
  • Graph DB schema (Neo4j) for lead graph

📐 Math Mentor AI

Personal Project

JEE-level multi-agent tutoring system with multimodal input and HITL feedback.

  • 6 agents: Memory · Parser · Router · Solver · Verifier · Explainer
  • Claude Vision OCR for handwritten problems
  • Whisper ASR for voice input
  • FAISS RAG + human-in-the-loop triggers
  • Streamlit UI · automated test suite

> experience_timeline

▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓
🚀 2025 – Present Founding ML Engineer
Visionary EdTech
Building the multilingual AI core — pure Python RAG for 12+ Indic languages, embedding optimization, hybrid retrieval at scale
🏢 2024 – Present AI/ML Engineer
Cyepro Solutions, Hyderabad
CRM-native lead scoring, automobile funnel analytics, codeblast analyzer with graph database design
🧑‍💻 2024 – Present Tech Lead
Infin AI Club, IIT Madras
Leading AI workshops, mentoring peers on ML projects, organizing hackathons and competitive ML events
🌾 2023 – 2024 AI/ML Engineer
Livestockify (AgriTech)
Vision-based poultry weight estimation, Wav2Vec2 audio classification (healthy vs sick chickens), IoT sensor fusion, investor-grade PDF reporting
🎯 2023 Organizer
Jarvis Club, IIT Madras
HackerRank placement prep contests · 100+ participants · end-to-end contest management
▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓

> tech_stack

Core Languages

AI / ML


Tools & Infrastructure


> featured_projects

🔬 Smart Poultry Farm Monitoring System

AgriTech · Computer Vision · Audio ML · IoT

End-to-end AI system for poultry health management combining vision, sound, and sensor data.

  • Wav2Vec2 fine-tuned for healthy vs. sick chicken audio classification
  • Vision pipeline for flock weight estimation (no physical contact)
  • IoT sensor fusion: temperature · ammonia · humidity
  • ReportLab investor-grade PDF health reports
  • Veterinary diagnostic report generation

🛰️ Satellite-Driven Aquaculture ML

Remote Sensing · Multi-task Learning

Water quality prediction for AP carp ponds using satellite imagery.

  • Sentinel-2 + Landsat-9 for DO, NH₃, pH, Chl-a prediction
  • LightGBM / XGBoost ensemble baseline
  • Late-fusion multi-task neural network
  • Backed by IIMA Ventures · IIT Kanpur TBI

📊 RAG Parameter Optimization Pipeline

Information Retrieval · NLP · Evaluation

GridSearchCV-driven systematic optimization of a RAG pipeline over CBSE Std 8 Science content.

  • Sweep over chunk size, overlap, embedding model, top-k
  • Metrics: MRR · NDCG · Hit@K
  • Multi-model embedding benchmarking
  • Reproducible Jupyter notebooks + embedding landscape guide

🚦 Smart Traffic Management — SIH

Computer Vision · Edge AI · Smart India Hackathon

AI-based adaptive traffic signal system for Indian urban intersections.

  • Real-time vehicle detection and density estimation
  • Dynamic signal timing based on live traffic load
  • Optimized for edge deployment constraints

> github_stats




> competitive_profiles


DSA · Problem Solving

ML Competitions

Competitive Coding

Algorithms

> what_I_believe

┌─────────────────────────────────────────────────────────────────────┐
│                                                                     │
│   "Most AI is built for Silicon Valley.                             │
│    I want to build AI for the farmer in Andhra Pradesh,             │
│    the student in rural Bihar, the shopkeeper in Tamil Nadu.        │
│                                                                     │
│    That means multilingual. That means edge-first.                  │
│    That means it actually has to work."                             │
│                                                                     │
│                                               — Sai Ruthvik         │
└─────────────────────────────────────────────────────────────────────┘

> let's_connect



 



📍  IIT Madras, Chennai / Hyderabad, India
🎓  B.S. Final Year — IIT Madras
⏰  IST (GMT+5:30) · Usually responds within 24h

contribution snake

Built from scratch · Not a template · Powered by curiosity ☕

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages