I like working close to the metal: file transfer protocols, intrusion detection on embedded networks, and retrieval pipelines that actually have to perform under load — not just demo well.

class Engineer:
    def __init__(self):
        self.focus = ["systems programming", "network security", "distributed systems"]
        self.currently_building = "JAVIN Share — cross-platform LAN file transfer"
        self.currently_learning = ["low-level networking internals", "DSA"]

What I'm building

• JAVIN Share — cross-platform LAN file transfer with no cloud dependency. Three interchangeable relay strategies (disk / stream / buffered) to trade off throughput, latency, and memory under different network conditions. Backpressure-aware buffering and sender-lock concurrency control to keep multi-GB transfers reliable over flaky Wi-Fi.

• Research RAG Assistant — Retrieval-Augmented Generation pipeline for grounded Q&A over research papers. Hierarchical chunking to preserve section structure, BGE embeddings indexed in FAISS, Gemini for context-grounded summarization.

• Hybrid CAN Bus Intrusion Detection System — real-time anomaly detection for in-vehicle networks, built around a Gaussian Naive Bayes classifier over engineered traffic features (timing, frequency, graph-relational). B.Tech research project, written up as a full academic report.

Stack

C++ Python Java JavaScript · Node.js Express Socket.IO · FAISS scikit-learn NetworkX · MySQL Firebase Git

Elsewhere