Sang Truong · Senior AI Engineer · Applied AI Lead

I build AI systems that have to work.

Senior AI engineer and technical lead with 6+ years across agentic AI, LLM/RAG systems, voice and real-time AI, computer vision and video intelligence, multimodal learning, quantitative ML, time-series forecasting, edge inference, and production AI engineering. I own systems end to end from research and data acquisition through model/tool design, retrieval and orchestration, APIs, distributed workloads, product surfaces, deployment, observability, reliability, and stakeholder delivery.

Ho Chi Minh City, Vietnam · GitHub @sangtrx

View flagship systems Technical depth Résumé
Current roleHead of Artificial IntelligenceApplied AI experience6+ yearsGraduate degreeMEng Computer Engineering · 4.0/4.0ResearchIJCV · AAAI Oral · IEEE JBHI · Poultry Science

Technical depth

One career, four deep engineering domains.

The current résumé family separates domain evidence instead of flattening everything into one generic skills list. The portfolio follows the same model: broad senior ownership, with inspectable depth in each domain.

01

Applied AI · Agents · Voice · RAG

Production AI systems where retrieval, tools, memory, voice/real-time interaction, authority, safety, evaluation, and distributed application state matter as much as the model call.

  • Ho Chi Minh City Traditional Medicine Hospital clinical AI: one semantic owner for request/reference/source/tool/scope, bounded approved-corpus research, durable Evidence Workspace, deterministic clinical authority and exact citations
  • FPT AI4U: Azure OpenAI, LangGraph/LangChain, Qdrant RAG, tool/model routing, code execution, web search, memory controls and guardrails
  • Voice and tool workflows: multilingual Azure Speech transcription/evaluation, Vietnamese ASR/TTS, WebSockets/SSE, Playwright-based browser verification and failure handling

FastAPI · LangGraph · LangChain · Azure OpenAI · Azure Speech · Qdrant · Milvus · pgvector · Playwright · Celery · RabbitMQ · Redis · PostgreSQL · Docker

02

Computer Vision · Video · Edge AI

Physical-world AI spanning camera/media ingest, perception, tracking, temporal events, evidence capture, edge inference, and recovery under unreliable real-world conditions.

  • Production multi-camera video intelligence: YOLO11, Vietnamese ALPR, ByteTrack, line crossing, event video, identity/freshness checks, watchdog recovery
  • University of Arkansas: temporal action understanding, vision-language modeling, industrial CV, CarcassFormer, YOLOv8 Jetson deployment
  • 5D Agriculture: autonomous braking, face recognition, Intel RealSense D435 RGB-D livestock measurement, embedded AI

PyTorch · TensorFlow · OpenCV · YOLO11/YOLOv8 · Fast-ALPR · ByteTrack · TensorRT · CUDA · ONNX Runtime · Jetson · FFmpeg · MediaMTX

03

Quantitative Research · Trading Systems

A research-to-production stack built around point-in-time evidence, reusable causal computation, leakage/multiplicity control, durable signal/risk state, execution/reconciliation, and public verification boundaries.

  • Curren research: Rust/Python causal core, Arrow/Parquet PIT evidence, shared timeframes/primitives/events, versioned event store, hypothesis views, global OOF and append-only Alpha History
  • Curren production path: normalized external alpha-source data, fail-closed ML quality gate, restart-safe lifecycle/risk, guarded execution, reconciliation and research↔streaming parity
  • Confidential Fund + Bluebelt: equity/crypto/FX quantitative research, ML ensembles, sentiment-derived signals, AWS execution and MLflow experimentation

Rust · Python · Arrow · PyArrow/Parquet · Polars · DuckDB · LightGBM · CatBoost · XGBoost · SciPy · Statsmodels · Optuna · NautilusTrader

04

Research · Multimodal & Temporal ML

Peer-reviewed research across temporal video understanding, vision-language learning, medical time-series representation learning, and industrial computer vision.

  • ABN → AEI → AOE-Net: action boundaries and actor/object/environment interaction modeling for long untrimmed video
  • VLCAP → VLTinT: contrastive vision-language learning and coherent video paragraph captioning; VLTinT was an AAAI 2023 Oral
  • sCL-ST + CarcassFormer: medical time-series contrastive learning and industrial localization/segmentation/classification

Transformers · Contrastive Learning · PyTorch · TensorFlow · Detectron2 · MATLAB · NumPy · SciPy · Scikit-learn · Weights & Biases

Deep case studies

Two systems with architecture and evidence exposed.

YHCT shows bounded clinical AI under a real hospital UAT boundary. Curren shows point-in-time quantitative research, live-system state, and a one-way public verification architecture. Both distinguish implementation, validation, deployment, and non-claims.

Case study2026

AI Architect / Lead Builder · EPIC TECHNOLOGY

Ho Chi Minh City Traditional Medicine Hospital

EPIC TECHNOLOGY · Clinical AI & Decision Support

A clinician-facing AI chatbot and decision-support platform with a single semantic-owner clinical-agent path, bounded approved-corpus evidence research, deterministic clinical fact authority, durable evidence state, exact citations, and explicit validation/production boundaries.

Single semantic ownerPostgreSQL evidence truthDeterministic clinical authority
Clinical AIAgentsFastAPINext.jsPostgreSQLQdrantRAG
Open case study
Case studyJun 2026 — Present

Independent AI/Quant Side Project · Solo Builder

Curren

Independent quantitative intelligence & trading-systems project

A solo-built multi-system platform spanning point-in-time research evidence, shared Rust/Python causal event infrastructure, OOF/multiplicity governance, real-time signal intelligence, fail-closed ML quality filtering, lifecycle/risk/execution, access/distribution, and an isolated public verification layer.

Shared causal event evidenceResearch↔streaming parityRead-only API · CLI · MCP
Quant ResearchRustPIT DataNautilusTraderMCPTrading Systems
Open case study

System portfolio

The work is broader than two case studies.

Selected systems across healthcare, education, enterprise agents, physical-world video AI, and quantitative research. Ownership labels are explicit so client/employer work is not presented as unrelated personal projects.

012026

EPIC TECHNOLOGY · AI Architect / Lead Builder

Ho Chi Minh City Traditional Medicine Hospital AI Chatbot & Clinical Decision-Support Platform

Clinical assistant for drug/herb lookup, interaction analysis, prescription review and governed knowledge Q&A using a single semantic-owner agent path, bounded local evidence research, deterministic clinical authority, durable observation/evidence state, exact citations and auditability.

Validated/deployed protected-UAT baseline · clinician-feedback stabilization in progress
022025 — Present

EPIC TECHNOLOGY · AI / Computer Vision Systems Lead

Production Multi-Camera Video Intelligence Platform

Real-time multi-camera system covering media acquisition, stream normalization, YOLO11 perception, Vietnamese ALPR, ByteTrack tracking, temporal event logic, evidence capture, alerts, access control, monitoring, and automated recovery.

Production physical-world AI
032025 — Present

EPIC TECHNOLOGY · AI Architect / Lead Builder

AI-Powered Open edX Platform

Teacher-reviewable course generation from textbooks/syllabi into lessons and OLX packages, plus Milvus-backed tutoring, secure question context and answer-key handling, Vietnamese ASR/TTS, WebSockets, Live2D browser interaction, Playwright verification, and Tutor/Docker/Nginx operations.

Education AI platform
042024 — 2025

FPT Software · AI Engineer

AI4U Enterprise Conversational Agent

Azure OpenAI agent with LangGraph/LangChain orchestration, Qdrant RAG, model/tool routing, controlled code/tool execution, web search, Mermaid generation, token-aware memory, content safety, multilingual speech/transcription evaluation, and distributed AI workloads.

Enterprise agent + voice platform
05Jun 2026 — Present

Independent side project · Solo Builder

Curren Quant Intelligence & Trading Systems

Causal research evidence, versioned event/hypothesis infrastructure, OOF/selection governance, signal intelligence, ML quality gating, durable lifecycle/risk/execution, reconciliation, access/distribution, and public API/CLI/MCP verification boundaries.

Research evidence → production state → public verification

End-to-end ownership

Beyond the model call.

The recurring pattern is turning research/model capability into a bounded system with explicit data, authority, failure, evaluation, operations, and acceptance contracts.

AI & system architecture

Translate ambiguous product goals into model/data contracts, authority boundaries, APIs, state, deployment, testing, observability, and acceptance criteria.

Model, tools & knowledge

Design agent orchestration, retrieval, tool/model routing, context and memory, browser/tool workflows, structured outputs, guardrails, provenance, and safe boundaries around side effects and fact ownership.

Backend & product engineering

Build the application path around the model: FastAPI services, durable state, async workers, streaming, integrations, product surfaces, permissions, and operational controls.

Evaluation & reliability

Define what “working” means, separate implemented from validated, test retrieval/response/speech and failure behavior, and make evidence visible instead of hiding uncertainty behind fluent output.

Cloud, edge & deployment

Carry systems through Docker/Linux, cloud and on-prem delivery, GPU/edge optimization, networking, observability, recovery, runtime verification, and stakeholder handoff.

Research to production

Move from experimental methods and reproducible evaluation into constrained production systems without erasing the assumptions that made the research valid.

Career

Research depth, product ownership, production constraints.

From embedded computer vision and quantitative systems to multimodal research, enterprise agents, hospital AI, education platforms, and production video intelligence.

Jun 2025 — Present

Head of Artificial Intelligence

EPIC TECHNOLOGY

Leading architecture and delivery across clinical AI, education AI, production computer vision/video intelligence, and AI platform engineering; owning requirements decomposition, data/model contracts, APIs, authority boundaries, deployment, testing, observability, and stakeholder acceptance.

Jun 2024 — Present

AI Tech Lead · Part-time / Consulting

A9 IOT

Leading real-time environmental IoT forecasting with continuous preprocessing, evaluation, inference, visualization, and stakeholder-facing reporting across LSTM, XGBoost, ARIMA, and Prophet workflows.

Oct 2024 — Jun 2025

AI Engineer

FPT Software

Built enterprise conversational agents and multilingual voice/real-time AI systems with Azure OpenAI, LangGraph/LangChain, Qdrant RAG, tool/model routing, distributed workers, guardrails, evaluation workflows, and production APIs.

May 2021 — Jan 2024

Graduate Researcher / PhD Student Researcher & Teaching Assistant

University of Arkansas

Graduate research in multimodal video understanding, temporal representation learning, vision-language modeling, medical time-series learning, industrial computer vision, and real-time edge inference; completed an MEng in Computer Engineering during this period.

Full chronology, teaching, skills and honors

Research

Peer-reviewed foundations.

Work across temporal video understanding, actor/object/environment interaction modeling, vision-language learning, medical time-series representation learning, and industrial computer vision. VLTinT was selected as an AAAI 2023 Oral.

Google Scholar
2024 CarcassFormer
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
Research rigor is only useful in production if the system preserves the contract that made the result valid.

In clinical AI that means explicit fact authority, provenance, and acceptance states. In quantitative systems it means point-in-time data, leakage controls, reproducible evaluation, and a clean boundary between research artifacts and live state.

Senior / lead roles

Need someone who can own the whole system?

I'm interested in applied AI leadership and senior engineering roles where model capability has to become reliable software: agent platforms, computer vision/video intelligence, multimodal AI, quantitative ML, distributed AI backends, and production delivery.

Start a conversation GitHub LinkedIn Full résumé