CS Undergraduate (9.55 CGPA) | Applied Computer Vision, Network Security & Robotics Building low-level systems, bridging them with machine learning, and containerizing them for edge deployment.
Stack: Python, Meta SAM 2, OpenCV, SQLite, Docker A containerized, hybrid computer vision pipeline integrating SAM 2 and OpenCV to handle foundation model failures.
- Built a deterministic logic layer to track distinct "islands" and maintain ID persistence when objects fragment in segmentation masks.
- Integrated an SQL telemetry backend for automated incident logging, engineered for edge-node deployment and observability.
Stack: Python, Scapy, Scikit-Learn (Random Forest), Docker A custom packet-sniffing engine built to categorize network traffic anomalies in real-time.
- Bridges low-level packet ingestion in promiscuous mode with a Random Forest classifier to identify normal vs. attack traffic.
- Deployed as an isolated Docker container leveraging Linux host-networking (
NET_RAW,NET_ADMIN) capabilities.
Stack: Python, Control Theory (PID/Kalman), Docker A 1D Digital Twin and Software-in-the-Loop (SITL) flight engine.
- Implements a recursive Bayesian filter (Kalman) for state estimation and a custom Domain-Specific Language (DSL) parser for mission logic.
- Containerized for headless execution to support parallelized Monte Carlo testing of flight control algorithms.
Stack: Python, Scikit-Learn, Pandas, Time-Series Analysis A predictive machine learning engine forecasting daily PM2.5 levels.
- Engineered lag and rolling features to capture temporal patterns from Central Pollution Control Board (CPCB) data.
- Automates classification of AQI levels to provide data-driven public health advisories.
Core Technologies
- Languages: Python, C/C++, Java, SQL, Bash
- Vision & AI: Meta SAM 2, OpenCV, Random Forest, Scikit-Learn
- Cloud-Native & DevOps: Docker, Linux/Unix internals, Git, SQLite
- Robotics & Control: ROS 2, Kalman Filters, PID, Bare-Metal C (STM32)
Current Engineering Focus & R&D:
- Cloud-Native & Distributed Systems: Transitioning standalone ML and robotics containers into distributed Kubernetes workloads (DaemonSets, Jobs) for scalable observability and Monte Carlo testing.
- Embedded Control: Porting software-in-the-loop flight algorithms to bare-metal C for STM32 (Cortex-M3) microcontrollers.
- Systems Architecture & Security: Deep-diving into OS internals, memory-optimized C data structures, and expanding real-time network packet inspection techniques via eBPF.