I build across the stack - from LLM agents and classic ML models to full-stack applications, search engines, backend microservices. I like shipping things that are useful, fast, and a little bit fun.

• I am an applied AI/ML Software Engineer with 2+ years of experience and 2 published research papers, who enjoys experimenting and loves building things.
• Formerly at Intel and Northeastern University, I have contributed to scalable full-stack web applications, LLM-RAG microservices, and humanoid agents saving 100,000$ annually.
• My personal projects span developing web/android applications, building NLP/CV models, creating Unity games, and even building my own search engine, cryptocurrency, compilers, and more!

• 🎓 MSCS @ Northeastern University
• 💼 Open to roles in AI/ML, backend/systems, or full-stack
• 📫 suria.abdul04@gmail.com · 🔗 LinkedIn: Abdulaziz Suria
• 🧰 Portfolio: https://abdulaziz-04.github.io

• Portfolio Analyzer (Java, Swing, MVC/MVVM, GraphQL/REST) – Flexible portfolios, plans, cost basis, caching, GUI, mock testing.
• Collaborative Painter (DLang, SDL, UDP/WebSockets) – Real-time multi-client canvas + chat, undo/redo, flood-fill speed-ups.
• OpsDesk (IT Logger) (React, Node, MongoDB, Docker, AWS EC2) – Filters, status chips, tech roster, light/dark, CRUD tickets.
• Tiltastic – Tile Visualizer (React, Flask, TensorFlow, AWS) – Semantic segmentation overlays for home decor; cached masks cut viz time from 60s → 30s.
• Accessible Drum Machine (React) – ARIA labels, keyboard mapping, screen-reader feedback, loop recorder.
• Shelfwise – Personal Library (React, Node, GraphQL) – Notes → BERT summaries, cosine-sim recommendations, per-book categories.

• LLM Story Generation (Falcon/LLaMA + LoRA/PEFT) – Genre-conditioned horror & fairy tales; evaluation via sentiment + LDA topics.
• Image Captioning (CNN-RNN & ViT-GPT2) – Comparative metrics and error modes.
• Brain Tumor Classification – Transfer learning + augmentation over 4 datasets.
• Mood Synergizer – Webcam emotion → curated music; DeepFace + desktop player.

• BitHeat – Blockchain in Node.js – PoW, difficulty retarget, ECDSA wallet, mempool, P2P WebSocket sync.
• Full-Stack Search Engine – Crawl → compressed inverted index → BM25/LM → PageRank/HITS → learning-to-rank → spam filtering.
• From Nand to Tetris – HDL chips → CPU → assembler → VM → Jack compiler in Java.
• Backend Microservices Suite (Node, Express) – URL shortener, timestamp, file metadata, browser whoami; production hardening notes.
• Price Notifier w/ WhatsApp (FastAPI, Twilio) – Checks Amazon prices on a schedule; sends alerts; Dockerized; CSV task list.

• Unity Gameplay Suite – Glitch Garden (lane defense), FPS Survivor (3 weapons + AI), Space Defender (SO waves), Block Breaker.
• Math Garden – Canvas handwriting + TF.js MLP digit recognition with playful growth feedback.
• Flutter Chat – Student chat with Firebase; AWS hosting.

• LLM-based Humanoid Agents in Social VR – CHI 2024 (memory-aware agents, GPT-4, VRChat).
• Statistical Analysis for TA Hiring – SIGCSE 2025 (Jensen-Shannon distance, t-tests/ANOVA, topic modeling).

Python   C#   JavaScript   Java   C++   SQL

React   Next.js   Tailwind   Bootstrap   HTML5   CSS3   Figma   Node.js   Spring Boot   Django   GraphQL   .NET

MySQL   PostgreSQL   MongoDB   Redis   Pinecone

AWS   Docker   Terraform   GitHub Actions   Apache Spark   Prometheus   Grafana

PyTorch   TensorFlow   Scikit-Learn   NumPy   Pandas   Matplotlib   Seaborn   OpenCV   NLTK   XGBoost   HuggingFace

Selenium   Elasticsearch   Git   Linux   Nginx   Jest   Cypress   Bash   Flutter   Unity