Hi! I'm Adrian Morton, a B.S. Computer Science student at Florida International University with a deep passion for full-stack engineering and designing complex systems. I'm an INIT Build Team Lead with a continuing drive to develop complex projects to deployment, working synchronously with my team to deliver quality solutions. Most of my work consists of backend engineering with the application of AI/ML technologies and frameworks, solving real-world problems with an efficient modern approach. Currently working on full-stack projects with AI/ML integration to simultaneously develop proficiency in both SWE and MLE.

Check out my site!

• INIT Build | Team Lead – Fall 2025

  • Architected the end-to-end ML system, from data ingestion and feature engineering pipelines to model training, evaluation, and deployment in production-like environments.​
  • Implemented and tuned advanced models (e.g., CatBoost, LightGBM, Temporal Fusion Transformer) with rigorous cross-validation and metric tracking to optimize performance on time-series workloads.
  • Owned core backend services and experimentation scripts in Python, enforcing code quality via Git workflows, modular design, and reproducible training runs.​

• INIT Build | Team Lead – Spring 2026

  • Leading the development of NoteBud, an AI-powered class notebook and study companion built with Python and Typescript that will use retrieval-augmented generation to answer student questions with citations from their own course materials.
  • Designing RAG pipeline using LlamaIndex/LangChain and Gemini to ingest PDFs/slides, chunk and embed content, process through retrieval ranking, groundedness scoring, and then return informed responses with highlighted evidence.
  • Developing a full-stack architecture with Next.js frontend, FastAPI backend, PostgreSQL + pgvector, Docker containerization, and Google Cloud object storage to manage users, classes, notebooks, and source files. ​

• AI4ALL | Fellow

  • Developed production-style ML models in Python (and C# where applicable), including Random Forest and XGBoost, to forecast real-world signals with high predictive accuracy.
  • Engineered features, performed hyperparameter tuning, and analyzed error patterns to improve model robustness, documenting pipelines and results for future iteration.​

Languages: Python • C/C++ • Java • SQL • TypeScript • JavaScript • Rust • Arduino

ML/AI: PyTorch • TensorFlow • LlamaIndex • LangChain • Google ADK • Qiskit • NumPy • pandas

Web & Backend: FastAPI • Next.js • React • Node.js • Flask • RESTful API • Docker

Databases: PostgreSQL • pgvector • Vector Databases • Apache Parquet

Cloud & Tools: Google Cloud • Git • Linux • Postman • Wireshark

AI-powered study companion with retrieval-augmented generation. Students ask questions and get answers cited from their own course materials.

• Stack: Next.js, FastAPI, PostgreSQL + pgvector, Docker, Google Cloud
• Features: RAG pipeline with LlamaIndex/Gemini, PDF ingestion, groundedness scoring, highlighted evidence

Predictive system for anticipating electric utility grid load peaks to enable dynamic pricing and resource allocation.

• Stack: PyTorch, CatBoost, LightGBM, Temporal Fusion Transformer
• Performance: R² = 0.9434 on TFT model

Agentic AI workbench that transforms founder ideas into structured business strategies through conversational intake.

• Stack: Google ADK, Gemini 2.5, FastAPI, LlamaIndex
• Features: Multi-agent system (onboarding, planning, research, analysis), session state management, RAG with vector search

High-throughput backend for utility grid operations with sub-minute situational awareness and hourly load forecasting.

• Stack: FastAPI, Apache Parquet, PyTorch, LightGBM
• Features: Automated retraining pipelines, 12-horizon forecasting, custom feature engineering