Final-year B.Tech CSE @ Lovely Professional University (CGPA: 8.03)
Data Science & ML Specialization · Minor: Financial Markets
20-week .NET Full Stack Training @ Capgemini · Based in Ongole, Andhra Pradesh 🇮🇳
I build production-grade full-stack applications that combine .NET microservices, Angular frontends, and Python ML pipelines. My projects span e-commerce intelligence, InsurTech, real-time financial systems, and computer vision — all with real metrics and deployed models.
- 🏦 Finance-aware engineer with a Minor in Financial Markets — built production FinTech systems (credit risk, real-time market intelligence, RAG pipelines) not just tutorials
- 🏗️ Comfortable with Clean Architecture, microservices, event-driven systems (Kafka), Docker, CI/CD
- 🤖 Shipped 10+ ML models in production — from XGBoost & LSTM to FinBERT and transfer-learning CNNs — across FinTech, e-commerce, and computer vision
- 📊 Data Science background with RAG pipelines, NLP, Computer Vision, and time-series forecasting
- 🔨 Currently building StreamPulse — a Kafka-based real-time financial platform, now in AWS deployment
- 🎯 Actively targeting roles in ML Engineering · AI/ML Full Stack · Data Science
Backend
Frontend
ML / AI
Data & Infrastructure
Apache Kafka · ASP.NET Core 10 · Angular 21 + PrimeNG Aura · Python FastAPI · TimescaleDB · Redis · SignalR · Prometheus + Grafana · AWS EC2
Event-driven platform processing live stock ticks through Kafka, with Python ML microservices for LSTM price forecasting, Isolation Forest anomaly detection, and news sentiment analysis — delivered to an Angular dashboard in real time over SignalR WebSockets, with full observability via Prometheus + Grafana.
- 🏗️ 12 services across .NET, Python & Angular, communicating through 4 Kafka topics
- 📈 Live candlestick charts, real-time ticker, and severity-based alerting with email notifications (Gmail SMTP via MailKit)
- ✅ 15 end-to-end integration tests (NUnit + FluentAssertions) tracing the full data pipeline
- 📊 Prometheus scraping all 7 services into a 7-panel Grafana monitoring dashboard
- ☁️ Currently deploying to AWS EC2, with GitHub Actions CI/CD up next
🛍️ ShopSense — AI-Powered E-Commerce Intelligence Platform ✅ Completed
Angular 21 · ASP.NET Core 10 (6 microservices) · 6 Python FastAPI ML services · Ocelot · Docker (17 containers) · SQL Server · Redis · RabbitMQ
21-day build. Full e-commerce platform with a complete AI intelligence layer — recommendations, dynamic pricing, churn prediction, sentiment analysis, demand forecasting, and fraud detection.
| Model | Algorithm | Metric |
|---|---|---|
| Sentiment Analysis | TF-IDF + Logistic Regression | 96.68% accuracy |
| Churn Prediction | XGBoost | ROC-AUC 0.9857 |
| Demand Forecasting | Holt-Winters | MAPE 42.9% |
| Dynamic Pricing | Gradient Boosting | R² 0.47 |
🤝 NexHire — AI-Powered HRMS
Angular 21 · ASP.NET Core 10 (3 microservices) · Groq API (Llama 3.1-8b) · Docker · GitHub Actions CI/CD
Built in 5 days for the FWC IT Services hackathon. AI-assisted recruitment, employee management, and performance evaluation with LLM integration via Groq.
🏥 SmartSure — Enterprise Insurance Management System
Angular 21 · ASP.NET Core 10 (4 microservices) · Clean Architecture · Razorpay · RabbitMQ · MailKit OTP · Docker
Production-grade InsurTech platform covering policy management, claims processing, fraud detection, and payment integration. Includes 117 NUnit tests with 90%+ coverage.
FastAPI · XGBoost · FinBERT · PyTorch LSTM · SHAP · Streamlit · Docker
253 features engineered from 307K+ loan records. 4-model calibrated ensemble with SHAP explainability, FinBERT news sentiment scoring, and LSTM 12-month default forecasting.
XGBoost ROC-AUC: 0.7875 · FinBERT F1: 98.45% · 🌐 Live Demo
Python · OpenCV · Scikit-learn · PyTorch (MobileNetV2 transfer learning) · LBP/GLCM texture features
4-class citrus quality-control pipeline (Greening, Canker, Black Spot vs. Fresh), built as a real assignment for Green Pod Labs. Two independent approaches — a handcrafted classical CV pipeline and a transfer-learning CNN — are trained and compared head-to-head on the same held-out test set, with pixel-level spoilage estimation and a packhouse-ready demo script.
- 🧪 Classical pipeline (HSV/Lab color + LBP/GLCM texture + Random Forest): macro-F1 0.954
- 🧠 MobileNetV2 + MLP head (frozen backbone): macro-F1 0.955
- 🩹 Pixel-level spoilage % estimation per defect class, validated with a healthy-fruit sanity check
- 📄 Full written report covering EDA, class-imbalance & background-bias mitigation, and deployment trade-offs
Anthropic Claude · HNSW Vector Indexing · Agentic Pipeline
RAG pipeline with HNSW vector search and an agentic Decompose → Retrieve → Synthesise workflow for intelligent financial document Q&A.
| Project | Stack | Highlight |
|---|---|---|
| 🤟 Indian Sign Language Translator | PyTorch · CNN · OpenCV | 5K+ gesture images, real-time inference <100ms |
| 🎬 YouTube Comment Sentiment | Python · BART · distilbart · YouTube API · Flask | Zero-shot NLP + summarization pipeline |
| 🏢 SQL Data Warehouse | SQL Server · ETL · Star Schema | Bronze-Silver-Gold medallion architecture |
| 📞 Telecom Churn Prediction | Python · Logistic Regression · K-Means | 5K+ records, 79.2% accuracy |
| 🏠 Airbnb Price Prediction | R · Random Forest | EDA + regression on listing features |
| 📈 Tableau Sales Dashboard | Tableau | Sales & customer KPI dashboards |
| 🛒 BigBasket EDA | Python · Pandas · Matplotlib | Category-level consumer insights |
| 🎓 B.Tech CSE | Lovely Professional University (2022–2026) · CGPA 7.92 · Data Science & ML Specialization (upGrad × LPU) · Minor: Financial Markets |
| 🏢 Full Stack .NET Training | Capgemini — 20-week enterprise developer program (Dec 2025–Present) |
| 📊 Data Analysis Training | Board Infinity |
| 🏆 Certifications | R Programming for Data Analytics (Board Infinity) · Hands-On PyTorch ML (LinkedIn Learning) |
Open to full-time roles in ML Engineering · AI/ML Full Stack · Data Science · Full Stack .NET