AI/ML Engineer Β· Computer Vision Β· LLM Applications Β· MLOps
Machine Learning Engineer with 5+ years of experience and a robust foundation in Software Engineering and Full-Stack Development. Specialized in decomposing complex problems into scalable, production-grade AI solutions. Proven expertise in architecture end-to-end ML pipelines, deploying Large Language Models (LLMs) and RAG systems, and orchestrating serverless workflows on AWS. Adept at bridging the gap between research models and reliable business applications.
- π§ AI/ML Engineer with 5+ years spanning Software Engineering, Full-Stack Development, and applied AI
- π MSc in Artificial Intelligence @ University of Science, VNU (2023β2026)
- Thesis: Extracting Information from Tires Using CNNs and Large Language Models
- π BSc in Software Engineering @ University of Science, VNU (2016β2021)
- π’ Currently Machine Learning Engineer at TecAlliance (promoted from SWE in 2022)
- π Based in Ho Chi Minh City, Vietnam
- π£οΈ Vietnamese (Native) Β· English (Professional Working Proficiency)
- π« thangnguyen31081998@gmail.com
Computer Vision ββββββββββββββββββββ Object detection (YOLO), image retrieval, OCR
Deep Learning ββββββββββββββββββ Model training, fine-tuning, CNN pipelines, PyTorch
MLOps & Infra βββββββββββββββ AWS Step Functions, CDK, Docker, CI/CD
LLM Applications ββββββββββββββββββββ Multi-agent orchestration, RAG, tool-using agents
Prompt Engineering ββββββββββββββββββ Structured outputs, few-shot retrieval, eval loops
Backend / APIs βββββββββββββββ FastAPI, microservices, REST API design
Languages
AI & Machine Learning
Web & APIs
Cloud, DevOps & MLOps
Databases & Vector Stores
π Tire Sidewall Extraction Β Β·Β MSc Thesis End-to-end CV pipeline extracting structured data from tire sidewall images.
- Boosted detection accuracy from 93.55% β 95.05% (@ IoU 0.5) by enhancing YOLOv11 with multi-branch features
- Geometric unwrapping + CLAHE preprocessing for low-light garage images
- Multi-LLM validation loop for robust JSON standardization
π Visual Search Engine Β Β·Β TecAlliance CNN-based car part image retrieval with scalable vector indexing.
- 88% top-k accuracy in similarity search; evaluated multiple CNN backbones
- Reduced manual part identification from hours to seconds
- Production vector index on Pinecone
π Vehicle Matching Pipeline Β Β·Β TecAlliance Serverless AWS pipeline reconciling New Zealand government schemas with internal datasets.
- 98% matching accuracy using decision trees + rule-based logic
- High-precision reconciliation across disparate schema sources
π Multi-Agent Extraction Platform Β Β·Β TecAlliance LLM-powered platform replacing manual internal document workflows.
- Significant reduction in operational workload across high-volume document pipelines
- FastAPI + React with Okta auth, multi-model orchestration, human-in-the-loop feedback
- Designed for external client expansion via REST API
π Multi-Agent Domain Translation Β Β·Β TecAlliance Domain-adaptive translation pipeline with autonomous quality control.
- Reduced LLM hallucination rates via LangChain multi-agent orchestration
- Knowledge-Graph-based n-gram index for terminological consistency
- Cycle Consistency validation loop for self-verifying outputs
π¬ RAG SQL Assistant Β Β·Β TecAlliance Natural language interface over databases and Azure Application Insights.
- Few-shot retrieval via ChromaDB + schema-aware tool usage
- Self-correcting execution loops eliminate ad-hoc engineering support for PO queries
π§° ML Developer Experience (DevEx) Β Β·Β TecAlliance Internal Python platform library standardizing ML development.
- Unified Okta auth + OCR/LLM integration across internal ML projects
- Embedded CloudWatch telemetry wrapper - traces performance and meters OCR/LLM cost per user and per project
βοΈ MLOps & Orchestration Β Β·Β TecAlliance End-to-end ML infrastructure on AWS.
- Multi-agent batch workflows in AWS Step Functions with automated retries
- Standardized Infrastructure-as-Code via AWS CDK
- CI/CD automation through Azure DevOps
- TecAlliance - Machine Learning Engineer (Oct 2022 β Present), Software Engineer (Nov 2021 β Oct 2022)
- Netwealth - Full-Stack Developer (Mar 2020 β Nov 2021).
- KΓ€rcher - Software Engineer Intern/Junior (Aug 2019 β Mar 2020).
- Professional Scrum Master I (PSM I)
- Professional Scrum Master II (PSM II)