Skip to content
View alok-devforge's full-sized avatar
🚀
🚀

Block or report alok-devforge

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
alok-devforge/README.md

Hey 👋, I'm Alok Kumar

tagline

Portfolio LinkedIn GitHub LeetCode Gmail Profile Views

B.Tech CST · IEM Kolkata ('27) · CGPA 8.71


4th-year Computer Science and Technology student and SDE Intern at Nevaeh Technology, working on core C++ modules for High-Frequency Trading (HFT) infrastructure with a focus on low-latency execution, performance, and system-level engineering.

I’m an AI-Native Developer who enjoys building end-to-end systems across C++, backend engineering, AI, and distributed architectures. I like working on challenging problems where performance, reliability, and practical product design matter.

A 5x Hackathon Winner, with recent back-to-back 1st-place finishes at Hack Synthesis 3.0 and Venomix 2026. I enjoy taking ideas from concept to working products under tight constraints.


◈ Experience

Software Development Engineer (SDE) Intern @ Nevaeh Technology (Aug 2026 – Present)

  • Engineering and optimizing core C++ modules for High-Frequency Trading (HFT) infrastructure.
  • Working on performance-critical components with a focus on ultra-low latency, reliability, and system-level optimization.

◈ Research

Machine Learning for Medical Image ClassificationProf. Dr. Subhalaxmi Chakraborty

  • Vision ensemble (85M+ params): EfficientNet-B4 + ViT-Base + ConvNeXt on 35K+ fundus images.
  • Focal Loss (γ=2.0) for class imbalance · ECE < 0.05 for publication-grade calibration.

◈ Projects

BhasaSetu — AI Vernacular Education PlatformReact Native TypeScript SQLite ONNX Kotlin
Offline-first education platform for tribal elementary classrooms · real-time Hindi/Bengali ↔ Santhali translation · multi-script support · on-device AI tutor · zero-internet classroom broadcast · CBT & gradebook system

CogniCodeNext.js FastAPI Groq LLM ChromaDB React Flow
AI codebase intelligence dashboard · sub-second semantic search across 10K+ files · 60% less API tokens · 35% faster refactoring

Rail SathiReact Native Node.js Socket.IO C++
Railway companion · GPS-speed locks (95%+ data integrity) · Smart Split-Seat Algorithm (+25% utilization) · IoT crowd detection (90% accuracy)

MAITRINode.js Socket.IO
Wildlife tracking · 1K+ concurrent streams at <50ms · Dynamic Safe Path algorithm (+40% pathfinding efficiency)


◈ Stack

Languages    C++ · C · Java · TypeScript · JavaScript · SQL
Frontend     React Native · Next.js · HTML5 · CSS3
Backend      Node.js · FastAPI · Socket.IO
Systems      HFT · Low-Latency Systems · System Design · Distributed Systems
Data         MySQL · SQLite · ChromaDB
AI/ML        Groq LLM · OpenCV · YOLO · ResNet · EfficientNet · ViT · ONNX
Tools        Git · GitHub

◈ Achievements

2026  ✦  1st Prize — Hack Synthesis 3.0
2026  ✦  1st Prize — Venomix 2026
2026  ✦  Winner (IoT) — Binary V2 Hackathon, KGEC Kalyani
2026  ✦  1st Runner-Up — Watch The Code, GEHU Haldwani
2025  ✦  1st Runner-Up — Technologia 1.0, UEM Kolkata
2026  ✦  Qualifier — Google The Big Code '26
2025  ✦  Finalist — IIT Madras Shaastra Coding Contest

Pinned Loading

  1. MediScan MediScan Public

    MediScan: AI-powered bone fracture detection system achieving 99.8% accuracy through deep learning. Features real-time X-ray analysis, transparent Grad-CAM visualizations, and clinical integration …

    Jupyter Notebook 41 7

  2. diabetic-retinopathy-detection diabetic-retinopathy-detection Public

    Multi-phase deep learning pipeline for diabetic retinopathy grading — ConvNeXt + Swin + EfficientNet ensemble, U-Net++ lesion segmentation, and XGBoost fusion achieving 0.92 QWK on 115K fundus images.

    Jupyter Notebook 1

  3. realtime-time-location-tracking-and-alert-system realtime-time-location-tracking-and-alert-system Public

    🌍 Real-time wildlife tracking & visitor monitoring system with GPS, WebSocket technology, proximity alerts, and interactive maps for University of Engineering and Management National Park

    CSS 4 1

  4. CogniCode CogniCode Public

    AI-powered codebase intelligence platform. Analyzes legacy codebases with real-time AST parsing, interactive dependency graphs, RAG-driven knowledge retrieval, and automated GitOps sandboxing.

    TypeScript 1 1