Systems-First Engineering
Backend Security & Systems Engineer
I build robust backend systems, think in networking fundamentals, live in Linux, explore cybersecurity, and use AI to accelerate — not replace — engineering depth.
Profile
I am a backend and systems engineer specializing in robust architecture, network performance, and cybersecurity protocols.
My approach is first-principles driven: I believe in understanding standard RFC specifications, parsing data streams, evaluating packet flows, and studying systems failure paths before writing code. I utilize AI tools and automated agents as execution multipliers to accelerate design cycles, but keep complete logical ownership of architecture.
Cybersecurity is a serious area of focus for me, especially network intrusion, Linux kernels hardening, and local protocol validation. I build code that holds up under load and prevents vectors of abuse.
Academics
Smt. Chandaben Mohanbhai Patel Institute of Computer Applications (CMPICA)
Graduated with a cumulative GPA of 8.25.
Computer Architecture, Operating Systems, Relational Databases, and Network Protocols.
Case Studies
An elegant, gesture-driven client built using Flutter and Dart for Same.Energy's semantic visual search engine. Replicates custom web-platform aesthetics in a highly-optimized mobile framework.
Translate a complex semantic image query web search into a physics-driven responsive mobile grid, handling asynchronous image pre-fetching and encrypted local caching on strict memory boundaries.
A full client integrating masonry visual grids, curated user feeds, and collections folders. It leverages physics-based scrolling and custom gestures with full-screen viewer overlays.
Decoupled Clean Architecture. Employs Riverpod for predictable reactive state boundaries, Dio client for network request intercepts/timeouts, GoRouter for type-safe navigations, and secure local keychain storage.
A secure, centralized, multi-factor attendance verification platform designed for academic institutions to eliminate proxy marking and location spoofing.
Built a high-performance backend using Go and Gin, leveraging PostgreSQL for durable records and Redis for fast transient session state. Facilitated real-time bidirectional communication via Gorilla WebSockets for live attendance dashboard updates.
Engineered a robust anti-spoofing pipeline combining rapidly rotating HMAC-SHA256 signed QR codes, on-device Google ML Kit face liveness checks (blink/smile detection), and BLE beacon proximity handshakes via a cross-platform Flutter application.
Implemented RS256 asymmetric JWT token signing, Argon2id password hashing, and mDNS/UDP LAN auto-discovery. Designed an audit-ready architecture with automated PDF/Excel report generation ensuring tamper-evident attendance validation.
A zero-cost, serverless, peer-to-peer secure file and text transfer utility. Connect directly browser-to-browser with no limits and no intermediate storage.
Leverages WebRTC data channels for high-bandwidth, direct connection streams between browsers. Implements public IP hashing with Mosquitto MQTT for seamless local network auto-discovery.
A Fabric mod for Minecraft Java 1.21 that provides in-game control for Spotify with overlay, HUD, and hotkeys. Communicates locally with a Spicetify bridge without needing API keys.
Built using the Fabric Modloader with a custom Java WebSocket client to securely communicate with the local Spicetify WebSocket server, providing real-time playback state sync.
Academic Publication
Presented at 7th International Conference on Soft Computing and Its Engineering Applications (icSoftComp 2025)
Abstract: Designed and deployed a lightweight, cooperative intrusion detection system (IDS) combining signature-based filtering with a quantized 3-layer Deep Neural Network (DNN) classifier. Optimized specifically to run inline on low-power ARM edge nodes, securing local edge topologies without compute bottlenecks.
| Evaluation Metric | Benchmarked Outcome |
|---|---|
| Detection Accuracy | 94.7% (Novel attack identification) |
| Low-Power Envelope | 4.2W per ARM active node |
| Packet Throughput | 15 kpps inline detection |
| Energy Efficiency | 67% reduction compared to standard x86 IDS |
Tools I Trust
Engineering Ethos
I believe in thinking from first principles — understanding data flow, control paths, and failure modes before writing a single line of code. I value performance, clarity, and maintainability over novelty. Given sufficient time and access, I focus on understanding the problem deeply and then building reliable, maintainable solutions.
Support
Open-source mods, research, and experiments — sponsorship and tips fund hosting and new features.
Collaboration
I'm always open to discussing new projects, backend systems, security research, internships, or engineering discussions.