📧 Email • 💼 LinkedIn • 🐙 GitHub • 📸 Photography

Fast Pass • Projects • Skills • Experience • Contact

TL;DR: CS grad with 4.0 GPA at GWU, graduating Dec 2025. I build production-ready ML/cloud apps — not just notebooks. Check the projects below, then let's talk.

What I bring:

• 5 shipped projects with real users/beta testers (InfraSight, TalkifyDocs, MelodyHub)
• Graduate TA for Design & Analysis of Algorithms — I know my theory and can explain it
• Full-stack ML engineer who can build the API, train the model, and deploy it

Core Stack:
Python · FastAPI · React/Next.js · TypeScript · PostgreSQL · MongoDB · Docker · AWS · scikit-learn · TensorFlow · LangChain

Actively seeking: Full-time Software Engineer / ML Engineer roles starting Jan 2026

📬 Best way to reach me: Email · LinkedIn

const abhiyan = {
  education: "M.S. Computer Science @ GWU (4.0 GPA)",
  graduating: "December 2025",
  role: "Graduate TA for Algorithms",
  passion: ["Building end-to-end ML apps", "Shipping real products", "Teaching complex concepts"],
  currentFocus: ["Cloud cost intelligence", "RAG-powered apps", "ML in production"],
  funFact: "I debug code and photograph the world 📸",
  openTo: "Full-time Software/ML Engineering roles (Jan 2026)"
};

What drives me:

• Taking algorithms from lectures and shipping them as actual applications with users
• Building end-to-end: data modeling → API design → ML pipelines → clean UIs
• Balancing deep technical work with product thinking

Beyond code:

• 👨‍🏫 Teaching algorithms to 60+ grad students
• 📸 Framing the world through a lens (see my work)
• 🎯 Shipping fast — iterate on live products, not perfect prototypes

💡 Tip: Screenshots coming soon! In the meantime, check out the repos for full docs, demos, and architecture details.
📌 Note: Don't forget to check my Pinned Repositories below for more projects!

InfraSight Cloud Cost Intelligence Platform Full-stack ML app that cuts AWS billing analysis time by 70%. Forecasting, anomaly detection, and cost clustering for DevOps/finance teams managing $10K+ monthly budgets. Impact: ✅ 92% prediction accuracy (Random Forest + LSTM) ✅ Catches cost spikes before billing cycles ✅ Enterprise-ready with Docker + JWT auth Tech: Python · FastAPI · PostgreSQL · scikit-learn · Docker · React · JWT

TalkifyDocs AI Document Assistant (SaaS) Enterprise RAG app that processes documents 47% faster than manual review. Chat with your PDFs using GPT-4, tested with 10 beta users from research firms. Impact: ✅ Sub-200ms query responses on 10GB+ repos ✅ 94% accuracy in document parsing ✅ Full SaaS: Auth, billing (Stripe), subscriptions Tech: Next.js · OpenAI GPT-4 · LangChain · Pinecone · Stripe · TypeScript · Prisma

MelodyHub Real-Time Social Music Platform Spotify-meets-Discord: synchronized music streaming with live chat. Supports 100+ concurrent users with 99.2% uptime during 2-month beta. Impact: ✅ Led 3-person Agile team ✅ Sub-3-second song loading (any network) ✅ OOP architecture → 60% faster feature dev Tech: Node.js · React · Socket.IO · Clerk Auth · Cloudinary · MongoDB

Disease Prediction ML Clinical Decision Support System Random Forest classifier on 4,920 medical records achieving 98.7% diagnostic accuracy. React UI designed for telemedicine deployment in rural Nepal. Impact: ✅ 131 symptoms, 41 disease classes ✅ Robust preprocessing pipeline ✅ Planned SHAP for explainable AI (FDA pathway) Tech: Python · scikit-learn · React · Tailwind CSS · Pandas · Jupyter

Audio Classification CNN Deep Learning for Animal Sounds Built CNN achieving 92% accuracy on dog/cat/bird audio clips. Explored CRNN, ViT, and transfer learning (YAMNet) architectures. Key Finding: Task-specific CNN trained from scratch outperformed transfer learning by 26% — showing pre-trained models aren't always better. Tech: TensorFlow · Keras · librosa · Mel-spectrograms · YAMNet · scikit-learn

Multi-Source Retrieval NLP Retrieval System RAG system integrating GPT-3.5 + FAISS that reduced research query times by 70%. Processes PDFs, YouTube transcripts, and audio files. Impact: Offline Mistral integration expanded reach by 40% for privacy-sensitive orgs (law firms, hospitals). Tech: Python · GPT-3.5-turbo · Mistral 7B · FAISS · LangChain · NLP

Design & Analysis of Algorithms (CSCI 6212) | Aug 2024 - Present | Washington, DC

• Teaching 60+ grad students dynamic programming, graph algorithms, and NP-completeness
• Designed interactive tutorials translating theory into industry case studies → 22% midterm score improvement
• Led TA training program: onboarded 8 new TAs, reduced training time from 2 weeks → 4 days

Fintech Platform Development | Jun 2023 - Sep 2023 | Lalitpur, Nepal

• Shipped 7 production features (React, Node.js, PostgreSQL) for 1,200+ active users
• Built automated booking system → eliminated 89% of scheduling conflicts (saved 15hrs/week for support)
• Led performance optimization → 43% load time improvement, conversion rates increased 12% → 16%

Digital Transformation & Infrastructure | Aug 2022 - Mar 2023 | Kathmandu, Nepal

• Rebuilt legacy website into lead-gen platform → 73% increase in qualified leads, $35K new contracts (Q1)
• Managed $12K annual IT budget as sole technical decision-maker → 40% operational cost reduction
• Trained 25-employee team, developed documentation → 80% reduction in external support tickets

George Washington University — Washington, DC
Master of Science in Computer Science | GPA: 4.0/4.0 | Jan 2024 - Dec 2025

Relevant Coursework:
Neural Networks & Deep Learning · Machine Learning · Cloud Computing · Software Engineering · Object-Oriented Design · Database Systems · Design & Analysis of Algorithms

St. Xavier's College — Kathmandu, Nepal
Bachelor of Science in Computer Science & Information Technology | GPA: 3.89/4.0 | Aug 2018 - Dec 2022

🔭 Currently working on:

• Graduating from GWU (Dec 2025)
• Teaching algorithms to 60+ graduate students
• Building ML-powered cloud applications

🌱 Currently learning:

• Advanced MLOps and model deployment patterns
• Kubernetes for container orchestration
• System design for ML at scale

💬 Ask me about:

• Building RAG applications from scratch
• ML model deployment and monitoring
• Full-stack development with React/FastAPI
• Teaching algorithms and data structures
• Photography and creative outlets for engineers

👯 Looking to collaborate on:

• Open-source ML/AI projects
• Cloud-native applications
• Production-ready RAG systems

When I'm not debugging or training models, I'm behind a camera capturing people, streets, and details most folks scroll past.

Think of it as debugging the real world — finding patterns in chaos, framing what matters.

Work I love:

• Turning messy requirements into clean APIs and useful dashboards
• Building ML-driven features: anomaly detection, recommendations, forecasting
• Optimizing systems: queries, pipelines, cost efficiency