Projects
A comprehensive portfolio spanning full-stack development, automation, data engineering, and financial technology.
Quantitative Trading Backtest System
Algorithmic Strategy Development Platform
A high-capacity algorithmic trading platform designed to validate custom Python strategies across the entire stock market. It leverages the Backtrader engine to perform large-scale backtesting on over 1000+ stock tickers simultaneously, automating the search for profitable strategy parameters.
- Massive Scale Testing: Engineered a data pipeline using Pandas and Backtrader to execute trading strategies against 1000+ historical stock datasets.
- Strategy Optimization: Implemented an automated parameter tuning module that iteratively tests strategy variables to generate optimal configuration sets.
- Web Management Interface: Developed a Django dashboard to orchestrate batch backtest runs and visualize comparative performance metrics.
Stock Market Analysis Platform
Desktop Investment Analytics Suite
A Tkinter-based desktop platform designed to help investors and traders analyze stock market data. It offers real-time data via yfinance, interactive Matplotlib charts, technical analysis tools, and algorithmic trading with custom strategies.
- GUI Design: Designed an intuitive GUI using Tkinter for enhanced user experience and complex data visualization.
- Data Integration: Integrated real-time market data fetching with yfinance to ensure up-to-the-minute accuracy for technical indicators.
- Algorithmic Trading: Developed custom algorithmic trading and backtesting functionality to simulate and validate investment strategies.
Kite Trade Mirroring System
Real-Time Order Replication Engine
A high-performance desktop application designed to replicate trading activity across multiple Zerodha Kite accounts in real-time. It acts as a bridge, listening for order updates from a 'Master' account and instantly executing identical orders in 'Child' accounts with millisecond latency.
- Event-Driven Architecture: Built using Twisted and Autobahn to handle asynchronous WebSocket streams, ensuring instant order detection.
- Desktop Interface: Developed a robust control panel using PyQt6, allowing users to map accounts and adjust risk management settings.
- API Integration: Integrated the Kite Connect API to manage order lifecycles, utilizing Pandas for position sizing calculations.
MetaTrader 4 Live Data Bridge
Financial Account Telemetry Pipeline
A low-latency financial data pipeline designed to bridge the MetaTrader 4 (MT4) terminal with modern web architectures. It functions as a real-time 'account logger,' instantly transmitting trade lifecycles and account health metrics to an external server for analysis and monitoring.
- FastAPI Backend: Engineered a high-performance FastAPI server to receive high-frequency webhooks from the trading terminal.
- MQL4 Integration: Developed a custom MQL4 Expert Advisor (EA) to extract and push live data points, including Open/Closed trades and P&L.
- Live Logger: Implemented a structured logging system to capture every account event, creating a granular audit trail.
IPO Market Intelligence Dashboard
Real-Time IPO Tracking Platform
A full-stack financial utility designed to aggregate and visualize comprehensive data for Initial Public Offerings (IPOs). It tracks current, upcoming, and listed IPOs, providing real-time insights into Grey Market Premium (GMP), subscription dates, and listing performance.
- High-Performance Backend: Engineered a concurrent backend using Go (Golang) to serve API requests and manage database connections.
- Data Aggregation Engine: Developed a multi-language scraping pipeline to extract live financial metrics from multiple stock market sources.
- Interactive Web Interface: Built a lightweight, responsive frontend using HTML/CSS/JS to display sorted IPO lists and detailed analytics.
Pujan-PET (Personal Expense Tracker)
Full-Stack Financial Management System
A comprehensive web application designed to track personal finances, manage recurring subscriptions, and visualize spending habits. It features a high-performance REST API backend and a responsive, data-rich frontend interface.
- Backend Architecture: Built a modular REST API using Python (FastAPI) to handle authentication and recurring transaction logic.
- Interactive UI: Developed a responsive frontend using Bootstrap 5 and integrated Chart.js for visual analytics.
- Data Management: Implemented AG Grid for advanced transaction filtering and PapaParse to enable bulk CSV imports.
Synthetic Financial Document Generator
OCR Training Data Generator
A Python-based automation tool designed to generate high-fidelity synthetic bank checks for training OCR and AI models. It addresses data scarcity in financial document processing by programmatically rendering realistic check layouts, MICR lines, and handwriting variations.
- Rendering Engine: Developed a dynamic rendering engine to create distinct check variations for robust model training.
- Typography: Integrated standard banking fonts (MICR E-13B) to ensure technical accuracy in generated datasets.
- Data Volume: Optimized the pipeline to produce high-volume, annotated image data for fine-tuning financial LLMs.
Payment Display Box
Embedded Real-Time Transaction Display
An embedded system that displays UPI payer name/ID and amount in real time, helping shopkeepers easily identify payers in crowded environments. Works alongside existing payment systems, ensuring no transactions are missed.
- System Design: Designed and developed a real-time UPI payment display system to function in high-volume retail settings.
- Visual Confirmation: Provided instant visual confirmation to prevent confusion or disputes, offering a modern alternative to sound boxes.
- Operational Efficiency: Improved shopkeeper efficiency by enabling quick payer identification during peak business hours.
Mobile Device Telemetry Agent
Remote System Administration Tool
A cross-platform mobile application developed to interface deeply with native iOS and Android APIs for remote system administration. It functions as a remote agent, capable of extracting granular device diagnostics, hardware identifiers, and local directory information for centralized monitoring.
- Native API Integration: Engineered a cross-platform mobile agent using React Native and Expo SDK to interface with low-level hardware APIs.
- Data Synchronization: Implemented secure data serialization pipelines to synchronize local device directories and diagnostics.
- Stream Optimization: Optimized data transmission efficiency using stream-buffers for reliable real-time telemetry reporting.
Forensic JSON Search Engine
Big Data Analysis Desktop Application
A specialized desktop application designed to index and search through massive JSON datasets (500GB+) with millisecond latency. It leverages the raw performance of Go for data processing while providing a modern, responsive GUI for analysts.
- High-Performance Core: Built using Wails (Go) to enable near-native speed for file I/O and memory management.
- Efficient Indexing: Developed a custom data ingestion pipeline to convert raw JSON into optimized binary formats (Parquet).
- Modern UI: Designed a responsive Vue.js frontend to visualize search results and filter data without freezing.
GSRTC Smart Transportation Ecosystem
Full-Stack Transportation Platform & Forensics Suite
A comprehensive engineering suite designed to modernize the bus travel experience for the Gujarat State Road Transport Corporation. It unifies a high-performance Next.js web platform, a Python-based desktop automation tool for real-time seat monitoring, and an AI-powered Flask backend.
- Full-Stack Architecture: Built a modern, SEO-optimized booking platform using Next.js (Frontend) and Flask (Backend).
- Automation & Forensics: Engineered a traffic interception engine using MITMProxy and Selenium to reverse-engineer private APIs.
- AI & Analytics: Integrated LangChain to allow natural language queries on transport data and developed a Tkinter desktop GUI.
Tor Proxy Load Balancer
High-Bandwidth Network Routing Utility
A sophisticated network utility designed to overcome the bandwidth limitations of the Tor network. It orchestrates a local pool of up to 32 concurrent Tor instances, aggregating their throughput to provide a high-speed, load-balanced proxy interface for web scraping or privacy-focused browsing.
- Parallel Architecture: Scripted a Python automation engine to initialize and manage 30+ isolated Tor processes.
- Process Management: Developed a custom Tkinter GUI to visualize the status of all instances and provide real-time controls.
- Throughput Optimization: Implemented a round-robin load balancing strategy to distribute outgoing network requests.
Music Studio Automation Pipeline
Reverse-Engineered Content Migration Tool
A complex Python automation suite designed to populate a custom music backend by bridging Spotify metadata with a private server. It automates the entire lifecycle of a song: link extraction, metadata scraping, file downloading, and uploading via reverse-engineered API endpoints.
- API Reverse Engineering: Analyzed and replicated private backend API calls to automate file uploads.
- Spotify Integration: Integrated spotipy and custom scrapers to fetch rich metadata, ensuring accurate mapping of Artists and Albums.
- Data Sanitation Engine: Developed a local JSON database system with maintenance scripts to normalize data and prevent duplicates.
Amazon Trend Analysis Tool
E-commerce Market Intelligence Scraper
A specialized Python automation tool designed to extract real-time market data from Amazon. It targets 'Trending' product categories to capture comprehensive metadata, enabling competitive analysis and market trend monitoring.
- Scraping Engine: Developed a robust scraping engine tailored to parse Amazon's complex DOM structure.
- Data Extraction: Automates the extraction of critical product metadata, including live pricing and customer ratings.
- Trend Logic: Engineered specific logic to identify and aggregate data from Amazon's trending and best-seller lists.
Universal OTT Media Extractor
Streaming Content Archival Suite
A specialized command-line automation suite designed to fetch, decrypt, and archive high-fidelity video content from streaming platforms like JioCinema and MX Player. It orchestrates N_m3u8DL-RE for multi-threaded HLS/DASH downloading and handles post-processing for a seamless viewing experience.
- Stream Automation: Scripted custom Batch workflows to automate the bulk downloading of entire series and episodes.
- DRM & Decryption: Integrated mp4decrypt binaries to handle protected content decryption pipelines efficiently.
- Media Processing: Automated FFmpeg operations to merge video, audio, and subtitle tracks into universal .mkv containers.
Vimeo Downloader
Cross-Platform Video Stream Extractor
A hybrid command-line utility designed to bypass content restrictions and download high-quality video streams from Vimeo. It utilizes a compiled Go core for rapid concurrent network requests and a Python wrapper for batch automation and file management.
- Core Logic (Go): Engineered a robust HTTP client in Go to parse master.json manifests and reconstruct fragmented video streams.
- Automation: Developed a Python automation wrapper to handle batch queueing, file naming, and directory organization.
- Architecture: Implemented a producer-consumer architecture to maximize bandwidth usage during multi-part downloads.
Site Ripper
Educational Course Archiver
A specialized web scraping framework designed to map, extract, and structure curriculum content from educational dashboards. It automates the retrieval of video links, metadata, and course hierarchies to build offline-viewable archives.
- Automation: Built a Selenium automation script to traverse complex DOM structures and extract deep-linked content.
- Traffic Interception: Implemented network traffic interception to capture hidden JSON payloads and authentication tokens.
- Data Parsing: Designed a parsing algorithm to convert raw HTML dashboard dumps into structured JSON course trees.
HEVC Optimizer
Automated Media Compression Pipeline
A media processing automation tool designed to drastically reduce video file storage requirements without compromising visual quality. It orchestrates FFmpeg to batch-convert large video libraries into the high-efficiency H.265 (HEVC) codec.
- Pipeline Automation: Scripted an automated pipeline using Python subprocess to manage complex FFmpeg encoding arguments.
- Storage Optimization: Optimized storage efficiency by achieving up to 50% file size reduction across large video archives.
- Reliability: Implemented error handling and logging to ensure reliability during long-running batch conversion tasks.
Synthetic Data Generation API (Faker)
DevTools & Testing Utility
A lightweight Python-based REST API designed to streamline the software testing lifecycle. It serves consistent, realistic mock data parsed from static datasets, enabling developers to validate frontend prototypes without reliable backend dependencies.
- API Development: Developed a custom Python API to parse and serve structured data from CSV files.
- Data Integrity: Engineered a simulation environment to validate data integrity before serving requests.
- Workflow Optimization: Optimized the development workflow by removing dependencies on live production databases for UI testing.