Skip to content
View SahilChambyal's full-sized avatar
🎯
Focusing
🎯
Focusing

Highlights

  • Pro

Organizations

@Quantrel-Labs

Block or report SahilChambyal

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

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

Report abuse
SahilChambyal/README.md

🚀 Top 0.5% in Amazon ML Challenge 2025 • 🎯 Microsoft Founder's Hub Member • 🔬 AI/ML @ Quantrel


Profile Views   Portfolio

🎓 B.Tech CSE @ Lovely Professional University (CGPA: 8.33)
💼 AI/ML @ Quantrel - Building next-gen LLM pipelines
🎖️ NCC Cadet - 6 years of service with A & B Certificates
📍 India | Open to Global Opportunities

💫 About Me

const sahil = {
    pronouns: "He/Him",
    role: "AI/ML Engineer",
    location: "India 🇮🇳",
    currentlyBuilding: "Agentic AI Systems & LLM Applications",
    education: {
        degree: "B.Tech in Computer Science",
        specialization: "Machine Learning & Deep Learning",
        university: "Lovely Professional University"
    },
    achievements: {
        amazonMLChallenge: "Top 0.5% (79th/20,000+ teams)",
        microsoftFoundersHub: "$20,000 in benefits",
        patentFiled: "Hybrid ViT Model for Crop Diagnosis"
    },
    motto: "Building AI that makes a difference 🚀"
};

🛠️ Tech Stack

🤖 AI/ML & Deep Learning

🔤 Languages

🧰 Frameworks & Tools

☁️ Cloud & Databases

⚙️ DevOps & Others


💼 Experience

🧠 AI/ML @ Quantrel

  • ⚡ Built high-throughput text-tokenization pipelines achieving 30-40% faster preprocessing
  • 🔍 Developed web-search tool for agentic models - 25% improvement in answer accuracy
  • 📊 Designed credit-consumption prediction algorithm with 90-95% cost-prediction accuracy

🚀 Featured Projects

🌾 Hybrid ViT for Crop Diagnosis

Patent Filed

  • Architecture: ViT + CNN + Custom MLP
  • Accuracy: 99.5% validation accuracy
  • Classes: 38 crop disease classifications
  • Feature: IoT-ready inference pipeline
  • Impact: Treatment recommendations included

🤖 ORIN Tri-Sense AI

Production

  • Type: Multilingual RAG Chat Runtime
  • Languages: 6+ supported
  • Performance: 200 QPS, <300ms latency
  • Relevance: 86% with 0.72 MRR
  • Impact: 96hrs → 10min response time

💰 DeBERTa Price Regressor

Kaggle

  • Model: DeBERTa-v3-base (183M params)
  • Dataset: 75,000 samples
  • SMAPE: 21.73% (from 63.17%)
  • Training: 30 epochs with AMP
  • GPU: Tesla P100-PCIE-16GB

🏆 Amazon ML Challenge 2025

Top 0.5%

  • Ranking: Top 0.5% globally
  • Position: 79th out of 20,000+ teams
  • Score: SMAPE 42.9%
  • Best Score: 39.2% (top participant)

📜 Certifications








🏆 Achievements & Recognition


79th / 20,000+ teams

Founder's Hub Member

6 Years of Service

Hybrid ViT Model

🤝 Let's Connect!



💭 "The best way to predict the future is to create it." — Peter Drucker


github contribution grid snake animation

⭐ From SahilChambyal with 💜

Pinned Loading

  1. Neural_Network_from_scratch_MNIST Neural_Network_from_scratch_MNIST Public

    Jupyter Notebook

  2. NLP_text_analysis_web-scrapping NLP_text_analysis_web-scrapping Public

    Jupyter Notebook

  3. GadgetRecommendationSystem-using-ML GadgetRecommendationSystem-using-ML Public

    A Machine Learning Project to help out with Gadget Recommendations

    Jupyter Notebook

  4. food-order-next-firebase-ts food-order-next-firebase-ts Public

    TypeScript 1

  5. portfolio portfolio Public

    running live

    TypeScript 1