"I build AI systems that don't just predict — they explain, scale, and create real-world impact."
I'm a B.Tech student in Artificial Intelligence & Machine Learning (GPA: 8.51/10) at NRI Institute of Technology, India — with production-grade ML experience across logistics, finance, and NLP.
- 🔬 Researcher in Explainable AI: Integrated Gradients, SHAP, Attention Rollout on transformer models
- 🏭 Industry Experience: Data Science @ AutonoPros — 95% accuracy price models, 20% revenue boost
- 🧑🏫 Recognized Top 50 AI/ML Mentor on Topmate — mentored 15+ learners
- 🌐 Open-source contributor — GirlScript Summer of Code
- ✍️ Technical blogger on Medium (300+ views per post)
- 📍 Hyderabad, India | Open to Remote Internships
| 🎯 Metric | 📊 Result |
|---|---|
| 🏆 Logistics Model Accuracy | 99.98% R² (MAE: 0.0021) |
| 🔍 Fraud Detection ROC-AUC | ≈ 0.95 (F1 > 0.94) |
| 🤖 Resume LLM Optimization | 78.1% avg keyword match improvement |
| 💸 AutonoPros Cost Reduction | 15% operational cost savings |
| 📈 AutonoPros Revenue Impact | 20% revenue boost via demand forecasting |
| 👥 Mentorship Reach | 15+ AI/ML professionals mentored |
Transformers are powerful — but why do they decide what they decide?
- Dataset: SST-2 Sentiment | Model: DistilBERT
- Implemented and rigorously benchmarked 3 XAI techniques:
- 🟣 Integrated Gradients — token attribution via gradient paths
- 🔵 Attention Rollout — propagated attention across transformer layers
- 🟡 SHAP — game-theoretic local explanations
- Evaluated stability, failure cases, and practical trade-offs for production ML systems
- Goal: Make LLMs debuggable and transparent for industry deployment
📦 Predictive Analysis of Delhivery Logistics — Team Lead | End-to-End ML Pipeline
Problem: Estimate delivery times accurately at scale for one of India's largest logistics companies.
Solution Stack:
- 🧹 Advanced EDA + feature engineering on large-scale operational data
- 🤖 Models: XGBoost, Random Forest, Gradient Boosting → Stacking Regressor
- 📊 MAE: 0.0021 | R²: 0.9998 — near-perfect predictive accuracy
- 🐳 Deployed via Flask/Django APIs + Docker + GitHub Actions CI/CD
Led a team of 5 from data pipeline design to production deployment.
🧾 AI-Powered Resume Optimization System — NLP + LLM Pipeline
Problem: Resumes fail ATS systems due to keyword mismatches — costing candidates opportunities.
Solution Stack:
- 🔎 TF-IDF vectorization + cosine similarity for gap detection
- 🌲 Random Forest classifier (72.84% acc) on 2,484 resumes across 24 job categories
- ✍️ GPT-4o-mini LLM rewriting → 78.1% average ATS keyword improvement
- End-to-end pipeline: parse → analyze → rewrite → score
💳 Credit Card Fraud Detection — Production-Grade ML System
Dataset: IEEE-CIS — 590K+ transactions, 400+ features
Solution Stack:
- 🔢 PCA dimensionality reduction → interpretable components (Transaction Velocity, Customer Risk Profile)
- ⚡ LightGBM classifier → F1 > 0.94 | ROC-AUC ≈ 0.95
- 🌐 Deployed as Streamlit web app with real-time + batch CSV prediction
- Production-ready: handles class imbalance, feature drift, and anomaly analysis
🏢 AutonoPros (Data Science Intern) Nov 2024 – Apr 2025 | Hyderabad, India
├── Price Prediction Model → 95% accuracy, -15% operational costs
├── Ride demand forecasting → +20% revenue boost
└── AWS + Docker ML pipelines → +30% workflow efficiency
🏢 Technocolabs Softwares (ML Engineer Intern) Nov – Dec 2024 | Indore, India
├── Logistics optimization → 90% prediction accuracy, -25% delivery delays
└── MLOps workflows → -50% data processing time
🏢 Outlier (Code Evaluator – Freelance) Nov 2024 | California, USA
└── 90%+ accuracy in LLM output code evaluations
Languages & Core
ML / AI Frameworks
Data & Visualization
Cloud & Deployment
| Certification | Issuer | Domain |
|---|---|---|
| Machine Learning Specialist | IBM | ML |
| Career Essentials in Generative AI | Microsoft | GenAI |
| MLOps with Generative AI | Google Cloud | MLOps |
| Azure AI Fundamentals | Microsoft | Cloud AI |
| ML Model Skill Badge | Google Cloud | Applied ML |
| PyTorch for Deep Learning | Infosys | Deep Learning |
| Business Analytics | Coursera | Analytics |
🔗 Full credentials: Credly · Google Cloud Skills · Microsoft Learn
- My Journey with AI: From Learning to Building Real-World Solutions — 300+ views
- Detailed Roadmap for ML Engineer in 2025 — 120 likes
- The Future of AI & ML — 150 likes
- Revolutionizing Data Science with ML for Predictive Analytics — 200 likes
Currently seeking: ML Research Internships · MS Admissions (USA) · Open-source Collaborations
"The best models aren't just accurate — they're explainable, deployable, and impactful."