class HarshitSingh:
def __init__(self):
self.name = "Harshit Singh"
self.college = "B.Tech @ NIT Calicut (Final Year, Batch 2027)"
self.location = "Varanasi, Uttar Pradesh 🇮🇳"
self.focus = ["GenAI Systems", "Agentic AI", "LLM Apps", "Backend Engineering"]
self.open_source = ["GSSoC'26 Contributor", "FOSSASIA Backend Developer"]
self.competitive = "LeetCode Guardian | Rating 2314 | Top 0.5% Globally"
self.experience = ["FOSSASIA", "GSSoC'26", "BHEL Intern", "Tathva Senior Tech Lead"]
self.motto = "Building systems that combine AI, engineering & real-world utility 🚀"
def say_hi(self):
print("Hi! Let's build something impactful together.")- 🔭 Currently building GenAI applications, LLM-based & Agentic AI architectures
- 🌍 Open Source — GSSoC'26 Mentee · FOSSASIA Backend Dev · eventyay (1.6k+ ⭐) · 1 PR Merged
- 🏆 LeetCode Guardian — Rating 2314 · Top 0.5% Globally
- 🚀 Senior Tech Lead @ Tathva · Tech Team @ Ragam · Past Intern @ BHEL
Production Django/Python platform · 1.6k+ ⭐ · Global GSoC org · PRs reviewed by Sourcery AI, GitHub Copilot & human maintainers
| PR | What I Fixed | Status |
|---|---|---|
| #3438 | Fixed typeahead filtering bug — prevented cross-event data leakage for limited team members | ✅ Merged |
| #3446 | Webhook reliability — 30s configurable timeout, break→continue fix, explicit error status |
🔄 Superseded |
| #3138 | Standardized login button text for i18n consistency | 🔄 Superseded |
| #3108 | First OSS contribution — login button text fix | 🔄 Superseded |
Autonomous AI agent that navigates, extracts & acts on the web end-to-end
- 🤖 Full agentic pipeline — LLM autonomously browses, reasons, and acts on web content
- 🔗 LangChain Agents + Tool Calling for structured multi-step web task execution
- 🧠 ReAct (Reason + Act) loop for intelligent decision-making across web sessions
- 📦 Modular architecture — extensible with new tools, APIs, and memory backends
Stack: Python LangChain Agents Web Automation LLM Tool Use REST APIs
Real-time AI news aggregator with smart summarization & topic analysis
- 📰 Live news from multiple sources processed with LLM-based summarization
- 🔍 NLP pipelines to categorize, tag, and extract key entities from articles
- 📊 Bias detection & sentiment analysis per news item for informed reading
- ⚡ Lightweight Python + Flask backend with clean REST API design
Stack: Python LangChain NLP Flask News APIs Summarization Sentiment Analysis
AI-powered medical assistant for symptom analysis & health insights
- 🩺 GenAI-powered health assistant that interprets symptoms and provides structured insights
- 📚 RAG architecture to ground responses in verified medical knowledge bases
- 🔒 Safe-response guardrails — outputs are informational, not prescriptive
- 🗄️ Vector DB-backed retrieval for fast, accurate, context-aware responses
Stack: Python RAG LangChain FAISS Vector DB Flask LLM Medical NLP
📂 Click to expand more projects
🤖 AI Simple Chatbot — Conversational LLM chatbot with prompt engineering & modular Python backend, extensible for RAG and multi-doc querying
📊 Smart AI Dashboard — Real-time AI-insight dashboard built with Streamlit; API-integrated, scalable, and focused on usability
🛒 Walmart Sales Analysis — End-to-end retail EDA with Pandas & NumPy; surfaces business trends and decision-support visualizations
Generative AI · RAG · NLP · Prompt Engineering · Vector Databases · Agentic AI · LLM Apps
⚡ Data Structures & Algorithms ████████████░░░ Active & Consistent | Guardian @ LC
🤖 Generative AI Systems ██████████░░░░░ Building Fast
🌐 Agentic AI & Web Agents █████████░░░░░░ Shipping Projects
🔗 LLM + RAG Architectures ████████░░░░░░░ Deep Diving
🌍 Open Source (FOSSASIA / GSSoC'26) ███████░░░░░░░░ Active Contributor
🔧 Backend + Microservices ██████░░░░░░░░░ Learning Phase
🗄️ Database Design (MongoDB/PostgreSQL) ███████░░░░░░░░ Advancing
"The best way to predict the future is to build it." — Alan Kay
