Skip to content
View algsoch's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report algsoch

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
algsoch/Readme.md

👨‍💻 Vicky Kumar

AI Engineer · Full-Stack Developer · Open-Source Builder

LinkedIn GitHub Kaggle Email

Building reliable, scalable, and production-ready systems across AI, backend engineering, and real-time applications.


🧭 About Me

I'm an AI Engineer and Full-Stack Developer passionate about building real-world systems that combine:

  • 🤖 Deep learning and LLM integration
  • 🏗️ Scalable backend architecture
  • ⚡ High-performance real-time applications
  • 🔧 Production-grade engineering practices

My work spans LLM evaluation, dataset creation, real-time communication engines, full-stack ML deployment, and open-source tooling. I focus on clarity, performance, and production reliability—shipping software that withstands real-world usage, not just demos.

🎓 Currently: Pursuing B.S. in Data Science at IIT Madras
🚀 Expanding into: Distributed systems, decentralized tech, and high-performance infrastructure


🛠️ Tech Stack & Engineering Strengths

🤖 AI/ML

  • CNNs, EfficientNet
  • Transformers, LLMs
  • RAG Pipelines
  • Fine-tuning (Phi-2, Whisper)
  • LangChain, Gemini API
  • PyTorch, TensorFlow

🔧 Backend & APIs

  • FastAPI, Spring Boot
  • Node.js, Express
  • REST, GraphQL
  • WebSockets, SSE
  • PostgreSQL, MongoDB
  • Redis, SQLAlchemy

🎨 Frontend & DevOps

  • React, TypeScript
  • Tailwind CSS, Vite
  • Docker, Kubernetes
  • GitHub Actions, CI/CD
  • Azure, Render, Vercel
  • Nginx, Load Balancing

🏗️ System Design Expertise

Distributed systems · Concurrency management · Latency reduction (<50ms) · Caching strategies · Performance tuning · Scalable architectures

Engineering Philosophy: Correctness, long-term maintainability, modularity, and real user-tested reliability.


🚀 Featured Projects

QuickPoll — Real-Time Voting Engine

High-performance WebSocket system handling 1000+ concurrent users with <50ms latency

  • ✅ Strict vote deduplication logic
  • ✅ 92% test coverage with comprehensive integration tests
  • ✅ Fully automated CI/CD deployments
  • �️ Stack: Node.js, WebSockets, Redis, Docker, GitHub Actions

🧠 Brain Tumor Detection — Full-Stack ML Platform

97.94% accurate CNN model deployed as production-ready web application

  • ✅ EfficientNet architecture with transfer learning
  • ✅ FastAPI backend + React frontend
  • ✅ Real-time prediction dashboards
  • ✅ Docker containerization + CI/CD pipeline
  • 🛠️ Stack: PyTorch, FastAPI, React, Docker, PostgreSQL

🤖 AI-Powered TDS Assistant — Multi-Capability AI System

All-in-one AI assistant for content analysis, automation, and workflow optimization

  • ✅ YouTube transcription with Whisper
  • ✅ PDF/Excel analysis and data extraction
  • ✅ GitHub API integration for automated workflows
  • ✅ CLI command execution and public API integration
  • 🛠️ Stack: Python, Whisper, LangChain, FastAPI, Gemini API

🎯 Dynamic Prompt Studio — AI Prompt Generation Platform

End-to-end AI tool for intelligent prompt engineering and content generation

  • ✅ Gemini API integration for LLM interactions
  • ✅ YouTube Data API for content analysis
  • ✅ Real-time analytics and metrics tracking
  • ✅ Context-aware prompt optimization
  • 🛠️ Stack: Python, Gemini API, FastAPI, React, WebSockets

Portfolio LLM Chatbot — Fine-Tuned Conversational AI

Custom-trained Phi-2 model for portfolio project retrieval and interaction

  • ✅ Fine-tuned on 40+ project descriptions
  • ✅ Context-aware retrieval with RAG pipeline
  • ✅ Custom React UI with real-time responses
  • ✅ Semantic search and intelligent filtering
  • 🛠️ Stack: Phi-2, LangChain, Transformers, FastAPI, React

🌱 Open-Source Contributions

🔗 Snapbit — First Open-Source Contribution

Significant improvements to mobile experience, authentication, and UI consistency

Contributions:

  • 📱 Enhanced mobile responsive design and touch interactions
  • 🔐 Improved authentication workflows and session management
  • ⚡ Optimized state handling and component performance
  • 🎨 Strengthened UI consistency across the platform

Impact: This contribution deepened my expertise in collaborative engineering, code review quality, and production-oriented frontend optimization.

🎯 Future Open-Source Goals

  • 🤖 AI/ML tooling and developer utilities
  • 🏗️ Distributed systems and infrastructure projects
  • 🔧 Stability-focused backend frameworks
  • 🌐 Decentralized application components

🕸️ Web3 & Decentralized Tech (Learning Journey)

Currently exploring:

  • ⛓️ Blockchain fundamentals and consensus mechanisms
  • 🏗️ Distributed architectures and protocol design
  • 🔮 Oracle patterns and off-chain computation
  • 🛡️ Protocol stability and security best practices

Leveraging existing strengths:

My foundation in APIs, concurrency management, containerized deployment, and real-time systems positions me well for building decentralized applications and off-chain services.


🏆 Professional Timeline

timeline
    title Career & Education Journey
    2024 : Started B.S. Data Science (IIT Madras)
         : Built QuickPoll Real-Time Voting Engine
         : Deployed Brain Tumor Detection ML Platform
         : Created HTML Citation Cleaner Tool
    2025 : AI-Powered TDS Assistant (Multi-capability AI)
         : Dynamic Prompt Studio (LLM Integration)
         : Portfolio LLM Chatbot (Fine-tuned Phi-2)
         : Mercor.io · Soul AI · Outlier (LLM Evaluation & Datasets)
         : First Open-Source Contribution (Snapbit)
         : Expanding into Decentralized Systems
Loading

📈 Key Milestones

Year Achievement Impact
2024 🎓 Enrolled at IIT Madras B.S. Data Science program
2024 ⚡ QuickPoll Launch 1000+ concurrent users, <50ms latency
2024 🧠 Brain Tumor Detection 97.94% accuracy, production deployment
2025 🤖 AI Assistant Suite 3 major AI-powered tools shipped
2025 💼 Industry Experience LLM evaluation, dataset creation, SFT
2025 🌱 Open-Source Start Contributing to production codebases

📄 Resume & Documentation

🤖 AI Engineer

Download

ML/DL · LLM · RAG

🔧 Full-Stack Developer

Download

Backend · Frontend · DevOps

📋 General Resume

Download

Complete Profile


📫 Let's Connect

LinkedIn GitHub Kaggle Email


🔥 Engineering Philosophy

"Ship fast. Think deeply. Build what lasts."

Every system should be reliable, understandable, and ready for scale.

┌─────────────────────────────────────────────────────────────┐
│  Correctness  →  Maintainability  →  Scalability  →  Impact │
└─────────────────────────────────────────────────────────────┘

💡 Open to collaboration on AI systems, distributed architecture, and production-grade tools

⭐ If you found my work interesting, consider starring my repositories!

Pinned Loading

  1. quickpoll quickpoll Public

    quickpoll challenge

    Python

  2. milestone-tracker milestone-tracker Public

    tracker

    JavaScript

  3. sachin sachin Public

    we have designed video portfolio

    CSS

  4. html-checker html-checker Public

    HTML checker is web app solution that help to clean my html code that contain unwanted text like cite_start or cite

    CSS

  5. Snapbit Snapbit Public

    Forked from yashudeveloper/Snapbit

    snapbit for habit building

    TypeScript