Huy Hoang — Software Engineer & AI/ML Researcher
Subject profile
Huy Hoang
Software engineer building reliable systems and agentic tools across Go, Python, cloud infrastructure, and applied ML. At Uber, I work on experimentation platforms, zero-downtime migrations, MCP integrations, and LLM-assisted review workflows.
University of Florida · B.S. in Computer Engineering · May 2027
▸
J.A.R.V.I.S.
▸ ARC REACTOR ........ ONLINE
▸ NEURAL INTERFACE .... SYNCED
▸ PORTFOLIO DATA ...... LOADED
0%
[ CLICK OR PRESS ANY KEY TO SKIP ]
Software engineer building reliable systems and agentic tools across Go, Python, cloud infrastructure, and applied ML. At Uber, I work on experimentation platforms, zero-downtime migrations, MCP integrations, and LLM-assisted review workflows.
Outside of work I mentor peers, lead engineering teams, and learn by building. I am especially interested in developer tools, production ML systems, and practical AI that makes complex workflows easier to understand.
Experience
Software Engineering Intern · Uber Technologies Inc.
May 2026 – July 2026
- Led search improvements project for an A/B experimentation platform end-to-end, implemented OpenSearch-native sort and filter on Go gRPC endpoints for 120k+ experiments with 27k+ searches/day
- Ran a zero-downtime schema migration in Go across protobuf, MySQL, Kafka, and OpenSearch indices, dual-writing on the live path and backfilling 120k+ production experiments through idempotent Cadence workflows at 99.8% completion
- Built a Go service exposing 7 experiment-management tools over Model Context Protocol (MCP), integrating it into an agentic debugging pipeline that root-causes misconfigured experiments for 1k+ services running experiments
- Architected a Python LLM reviewer for A/B experiment configs across 500 diffs/day, encoding past-incident patterns as lazy-loaded rule packs that cut agent context 50%, and built the eval harness measuring 94% precision and 86% recall
Machine Learning Research Assistant · University of Florida – Agriculture Precision Lab
Sep 2024 – Sep 2025
- Developed GPU-accelerated PyTorch pipelines for dense feature matching, increasing match reliability by 24% in dense foliage and improving multi-view tracking stability
- Architected a pipeline to consolidate 1K+ per-frame multi-view COLMAP reconstructions into a unified 3D field-scale model
- Applied monocular depth estimation to densify point clouds, achieving 90%+ data completeness for volumetric analysis
Software Engineer Intern · FPT Corporation
May 2025 – Aug 2025
- Implemented RESTful APIs in Java Spring Boot with WSO2 mediation policies, deployed with AWS EC2, Docker, and monitored via Grafana and Prometheus, achieving 98% request accuracy and improving response time by 20ms
- Configured and integrated 12+ enterprise microservices using WSO2 API Gateway deployed on Kubernetes, reducing cross-service failures by 20% and increasing system resilience in production clusters
- Developed custom Java WSO2 handlers for payload encryption/decryption and token validation using JWT and OAuth2, improving gateway throughput by 25% and enforcing secure, authenticated API communication across teams
Site Reliability Engineer Intern · FPT Information System
May 2024 – Aug 2024
- Automated AWS EKS infrastructure provisioning with Terraform and GitOps, improving deployment efficiency by 30%
- Built CI/CD pipelines with Jenkins and Docker, enabling instant rollbacks and cutting deployment time by 90%
- Deployed Prometheus monitoring across 20+ EC2 instances, cutting network setup from 2h to 15min and crash rates by 30%
Teaching Assistant · University of Florida (Programming, DSA, Algorithm Abstract, Dlog, Databases)
Aug 2024 – Present
- Led mini-lectures and structured office hours for 50+ students; section grades averaged 15% higher than other section
Projects
Horus’ Eyes AI Platform
🏆 1st Place — Best Use of Terraform (SwampHacks)
AI assistive platform for visually impaired users, with automated cloud infrastructure and a low-latency real-time inference pipeline.
Terraform · AWS · YOLOv8 · Flask · Docker · AWS Bedrock
Context-Aware LLM Application
Generative AI application with optimized state and memory management for accurate multi-turn conversations without latency degradation.
LangChain · AWS Bedrock · Python · FastAPI · Docker
Crave
Native social food discovery prototype for sharing restaurant visits, dish ratings, and cuisine rankings with ordering, reservations, and ride integrations.
SwiftUI · Kotlin · Jetpack Compose · Material 3
Perch
Personal AI IDE and agent app with a WebSocket server, headless Claude Code runner, terminal manager, and shared web/desktop client architecture.
TypeScript · Node.js · WebSockets · SQLite · Rust
Hower
Time management system that helps users track and manage how they spend their time through a Flask backend and Electron desktop application.
Python · Flask · SQLAlchemy · Electron
Search & Rescue AI
Mobile app with voice-activated incident reporting, GPS tracking, and real-time dispatch coordination.
React Native · FastAPI · AWS DynamoDB · S3 · Whisper
Skills
LANGUAGES
Python · Go · C++ · Java · Shell · SQL · Protobuf · TypeScript · JavaScript
FRAMEWORKS
FastAPI · Flask · Spring · Node.js · React Native · React
DEVELOPER TOOLS
AWS · Terraform · Kubernetes · Kafka · MySQL · Docker · Jenkins · OpenSearch · Cadence
AI/ML
Model Context Protocol (MCP) · LangChain · PyTorch · TensorFlow
LIBRARIES
OpenCV · NumPy · pandas · COLMAP
CERTIFICATIONS
NVIDIA Fundamentals of Deep Learning
Leadership
Software Team Lead · SASE Engineering Team
Sep 2023 – Apr 2024
- Led fine-tuning on custom audio datasets to improve scream detection accuracy
- Built Python audio source localizer via multi-mic time-of-arrival analysis