About
Research-trained. Systems-minded. Production accountable.
I'm Sang Truong (@sangtrx), a senior AI engineer based in Ho Chi Minh City, Vietnam. My work spans model research, LLM/agent systems, computer vision and video intelligence, quantitative ML, distributed backends, product engineering, cloud/edge deployment, and the operational boundaries that decide whether an AI system is actually trustworthy.
Operating model
Own the path from research to acceptance.
I am most useful when a system crosses several disciplines at once: research assumptions, data contracts, model/tool behavior, backend state, user experience, infrastructure, and stakeholder acceptance. I do not treat the model as the whole product.
My current Head of AI role covers clinical AI, education AI, production computer vision/video intelligence, and AI platform engineering. Earlier work spans enterprise conversational agents, multilingual speech systems, environmental forecasting, quantitative research, multimodal/video research, and embedded computer vision.
I prefer explicit boundaries over architecture theater: which component owns facts, what data was available at decision time, what can fail, what is recoverable, what has actually been validated, and what the system deliberately does not claim.
Domain depth
Broad systems ownership, with real depth where it matters.
The résumé keeps one canonical chronology while expanding evidence by domain. These are the four recurring technical centers.
Applied AI · Agents · Voice · RAG
Production AI where knowledge, tools, state, safety, evaluation, and user workflows matter as much as the model call.
- Clinical decision support with governed knowledge, citations, deterministic application authority, and explicit human/production boundaries
- Enterprise conversational AI with retrieval, tool/model routing, memory controls, guardrails, and distributed workloads
- Multilingual speech and real-time interaction across API, worker, and browser-facing paths
Python · FastAPI · LangGraph/LangChain · Azure OpenAI · Qdrant/Milvus/pgvector · PostgreSQL · Playwright · Celery/RabbitMQ/Redis
Computer Vision · Video · Edge AI
Physical-world AI spanning camera/media input, perception, tracking, temporal state, evidence, target-device inference, and recovery.
- Production multi-camera video intelligence across heterogeneous real-world sources and operational failure modes
- Industrial and academic vision workflows for inspection, recognition, tracking, and workstation state
- Edge optimization with TensorRT/CUDA/Jetson plus RGB-D and embedded perception work
PyTorch · TensorFlow · OpenCV · YOLO · ALPR · Tracking · TensorRT · CUDA · ONNX Runtime · Jetson · RTSP/HLS · FFmpeg
Quantitative Research · Trading Systems
Research and runtime engineering built around point-in-time evidence, causal computation, disciplined evaluation, durable state, and guarded execution boundaries.
- Curren: point-in-time research pipeline and shared research/runtime semantics without exposing strategy internals
- Leakage-aware evaluation, reproducible data artifacts, selection controls, lifecycle/risk state, and reconciliation
- Earlier equity, crypto, and FX research with ML ensembles, market data, cloud execution, and experiment tracking
Rust · Python · Arrow/Parquet · Polars · DuckDB · LightGBM · CatBoost · XGBoost · Scikit-learn · SciPy · Statsmodels · Optuna · NautilusTrader
Research · Multimodal & Temporal ML
Peer-reviewed research across temporal video understanding, vision-language learning, medical time-series representation learning, and industrial computer vision.
- Temporal action proposal generation and interaction modeling for long untrimmed video
- Vision-language learning and coherent video paragraph captioning, including an AAAI 2023 Oral
- Medical time-series representation learning and industrial visual inspection research
Transformers · Contrastive Learning · PyTorch · TensorFlow · Detectron2 · MATLAB · NumPy · SciPy · Scikit-learn · Weights & Biases
How I think
Good AI systems are contracts, not demos.
Give models room to reason where useful while keeping authoritative facts, permissions, tools, and failure behavior explicit.
In research and live systems, respect what information existed at decision time. Point-in-time discipline and provenance are engineering requirements.
Separate fluent output from the source, tool result, experiment, runtime observation, or acceptance gate that supports it.
State, retries, idempotency, restart behavior, monitoring, recovery, access control, deployment and handoff are part of the product.
For physical-world AI, camera/media reliability and target-device inference behavior matter more than isolated benchmark numbers.
Implemented, tested, deployed, accepted, and production-ready are different claims. I keep them different.
Technical stack
Tools organized by system responsibility.
The stack is broad because the systems cross model, data, product, infrastructure, and physical-world boundaries.
Applied AI / LLM / agents
LLM/RAG/agents · LangGraph/LangChain · tool/model routing · memory/context · grounded generation · guardrails · human-review boundaries · evaluation
Retrieval & knowledge systems
Qdrant · Milvus · Pinecone · pgvector · PostgreSQL · embeddings · governed ingestion · metadata filtering · citations · provenance
Voice / real-time AI
Azure Speech · multilingual ASR · Vietnamese ASR/TTS · speech evaluation · WebSockets/SSE · AsyncIO · background processing
Computer vision / video / edge
PyTorch · TensorFlow · OpenCV · YOLO · Detectron2 · ALPR · tracking · TensorRT · ONNX Runtime · CUDA · Jetson · RGB-D · RTSP/HLS · FFmpeg
Quantitative ML & research data
Rust · Python · point-in-time data · Arrow/Parquet · Polars · DuckDB · LightGBM · CatBoost · XGBoost · Scikit-learn · SciPy · Statsmodels · Optuna · NautilusTrader
Backend & distributed systems
Python · FastAPI · REST/SSE · WebSockets · Celery · RabbitMQ · Redis · PostgreSQL · SQLite · MongoDB · MinIO/S3 · retries · idempotency · durable state
Product, cloud & delivery
Next.js · React · TypeScript/JavaScript · Open edX · Playwright · Docker · Kubernetes · Linux · systemd · Nginx · AWS · Azure
Languages
English — professional working proficiency · Vietnamese — native
Research trajectory
From temporal video reasoning to applied AI systems.
Temporal action understanding
ABN → AEI → AOE-Net: action-boundary and interaction modeling for long untrimmed video, culminating in IJCV.
Vision-language video understanding
VLCAP → VLTinT: contrastive and Transformer-based modeling for coherent video paragraph captioning; VLTinT was selected as an AAAI 2023 Oral.
Medical & industrial ML
Medical time-series representation learning plus industrial/poultry vision and target-hardware edge optimization.
Education
Research-trained engineering.
Master of Science in Computer Engineering (MSCmpE) · University of Arkansas
GPA 4.0/4.0 · thesis: “Towards Multi-modal Interpretable Video Understanding” · advisor: Prof. Ngan Le.
Bachelor of Engineering in Automation and Control Engineering · International University — VNU HCMC
GPA 3.50/4.00 · Very Good · Top 1% · half-tuition scholarship recipient.
Selected research
Peer-reviewed work across multiple AI domains.
Simultaneous localization, segmentation, and classification of poultry carcass defects Poultry Science 2023 AOE-Net
Entity-interaction modeling with adaptive attention for temporal action proposal generation International Journal of Computer Vision 2023 VLTinT
Visual-linguistic Transformer-in-Transformer for coherent video paragraph captioning · AAAI Oral AAAI 2023 sCL-ST
Supervised contrastive learning with semantic transformations for multi-lead ECG arrhythmia classification IEEE JBHI 2022 VLCAP
Vision-language contrastive learning for coherent video paragraph captioning IEEE ICIP 2021 AEI
Actor-environment interaction with adaptive attention for temporal action proposal generation BMVC 2021 ABN
Agent-aware boundary networks for temporal action proposal generation IEEE Access 2021 Multi-module RCNN + Transformer
Multi-module recurrent convolutional neural network with Transformer encoder for ECG arrhythmia classification IEEE BHI Full Google Scholar profile
Explore