Sahil Belchada
Bridging Intelligence & Data Infrastructure
A quick glimpse into my architectural focus and problem-solving mindset.
AI & RAG Systems
Specialized in production LLM integrations (Gemini, Groq), LangChain workflows, and ChromaDB vector stores with offline-resilient fallbacks (Faster-Whisper) for mission-critical apps.
Real-Time Data Pipelines
Architect of high-throughput Extract-Validate-Clean-Load (ETL) pipelines using Apache Kafka event streaming, Apache Airflow DAGs, and dimensional star-schema PostgreSQL warehouses.
Scalable Microservices
Experienced building polyglot backends with FastAPI (Python) and Spring Boot (Java), Redis caching, Docker containerization, and real-time WebSocket communication.
Featured Projects
Real-world production systems, platforms, and interactive applications.
MarketMind AI
Real-Time Financial Analytics & Market Prediction Platform
A high-performance microservices architecture powering real-time financial market analytics, ML forecasting, and automated ETL pipelines:
- Microservices Engine: Spring Boot (Java) ingestion service + FastAPI (Python) analytics service connected via Apache Kafka streaming.
- Data Ingestion: Ingests live prices every 30s from Yahoo Finance API with Kafka producer-consumer streaming and broker-failure fallback.
- ETL & Warehousing: Orchestrated with Apache Airflow DAGs feeding a dimensional star-schema PostgreSQL warehouse.
- ML Forecasting & Risk: Linear Regression, Prophet, LSTM models with a model registry (RMSE, MAPE, R²) and financial risk metrics (Beta, Sharpe, VaR).
- Real-Time & Monitoring: Redis query cache, WebSocket live price feeds, Prometheus/Grafana metrics, containerized with Docker Compose.
HaritKranti
Farmers' Marketplace & Real-Time Crop Guidance Platform
A comprehensive MERN stack platform designed to empower agricultural communities with fair trade and technical advisory:
- Marketplace & Guidance: Direct product listing hub with dynamic crop management advisories and disease prevention guides.
- Multilingual Engine: Real-time translation API integration enabling seamless UI localization into multiple regional Indian languages.
- Robust Architecture: Node.js and Express REST APIs with MongoDB Atlas and optimized indexing schemas for fast queries.
- Security: Secure JWT-based authentication and role-based access for farmers and buyers.
MoM Generator
AI-Powered Meeting Minutes, Transcription & RAG Copilot
Intelligent enterprise transcription and summarization platform engineered during internship at the Software Development Cell:
- RAG Chatbot: Built ChromaDB vector store and LangChain agent in Python with robust fallback routing to prevent timeouts.
- Offline Whisper Fallback: Local Faster-Whisper pipeline in Python for resilient speech-to-text without network dependency.
- LLM Pipeline: Integrated Gemini & Groq for structured action item extraction and executive summary generation.
- Audio Chunk Merging: Programmed client-side native mono WAV chunk merging in Dart/Flutter for speaker diarization.
- Secure Backend: FastAPI with psycopg2/PostgreSQL, client disconnection monitors, JWT authentication, and Docker deployment.
Aurizen
Social Habit Tracking App with Offline-First Architecture
A modern habit formation platform engineered with offline-first synchronization and real-time social accountability:
- Offline-First SQLite: Built local SQLite database caching with automated bidirectional sync and conflict resolution.
- Supabase Cloud: Utilized PostgreSQL, Row Level Security (RLS) policies, and Realtime WebSocket subscriptions.
- Live Notifications: Firebase Cloud Messaging (FCM) integration for habit reminders and social cheer feeds.
AI Personal Finance Tracker
Automated Bank CSV Ingestion & Spending Insights
Full-stack personal finance application eliminating manual entry through automated data pipelines and AI spending analysis:
- Automated Ingestion ETL: Ingests raw bank statement CSVs, cleans and deduplicates records, and auto-categorizes transactions.
- AI Financial Copilot: Synthesizes monthly spending habits to suggest actionable budget savings.
- Stack: Built with Next.js modern app router, Node.js API routes, and MongoDB aggregation pipelines.
Technical Constellation
Categorized competencies across AI engineering, distributed data pipelines, and full-stack development.
AI & Machine Learning
Data Pipelines & ETL
Backend & Microservices
Databases & Caching
Cloud, DevOps & Monitoring
Languages & Frontend
Experience & Education
Real-world institutional development experience, academic foundation, and certifications.
Flutter & Backend Developer Intern
Software Development Cell (SwDC) • K.J. Somaiya School of Engineering
Lead Project: MoM Generator — Enterprise AI Meeting Minutes & Audio Transcription Tool
- Built a RAG-based chatbot vector store in Python using ChromaDB and LangChain with fast-fallback routing logic.
- Implemented offline, local Faster-Whisper speech-to-text transcription fallback without internet dependence.
- Built FastAPI backend services with psycopg2 for PostgreSQL integration handling storage and meeting record normalization.
- Integrated Gemini API and Groq for automated summarization and NLP-driven action item extraction.
- Programmed client-side native mono WAV chunk merging in Flutter/Dart to empower speaker diarization.
- Optimized CPU/GPU resource usage via custom client disconnection monitors in main.py, containerizing with Docker Compose.
B.Tech in Computer Engineering
K.J. Somaiya College of Engineering, Mumbai
Focused coursework in Data Structures & Algorithms, Distributed Systems, Database Management Systems, Computer Networks, and Artificial Intelligence.
Professional Certifications
Cloud Architecture & Product Design
Let's Build Something Extraordinary
Looking for an AI engineer, backend architect, or versatile builder? Let's connect.
Get in Touch
Feel free to reach out directly via email, phone, or connect on GitHub. I respond quickly to discussions about exciting engineering challenges.