Senior AI Engineer

About Axtra Labs

Axtra Labs is a Vietnam-based engineering company that works as a long-term technical partner for international clients. We build and deliver software and hardware-adjacent engineering projects end to end, from early prototypes to production systems.

Location: Remote, open to candidates based in Vietnam only.

About the role

We are looking for a Senior AI Engineer to design, build and ship AI systems for our international clients: LLM applications (RAG pipelines and agent-based systems), Computer Vision models, and optimized inference on edge devices. You will work directly with client stakeholders and own technical decisions from prototype to production.

Key responsibilities

- Design, build and deploy LLM-powered applications, including RAG pipelines, tool-using agents and evaluation frameworks

- Develop and train Computer Vision models for detection, segmentation, classification and tracking

- Optimize models for edge deployment (quantization, pruning, ONNX/TensorRT) on NVIDIA Jetson and similar hardware

- Build data pipelines for dataset collection, labeling, augmentation and versioning

- Set up MLOps practices: experiment tracking, model registry, CI/CD for models, monitoring and retraining

- Benchmark accuracy, latency, throughput and cost, and recommend clear trade-offs

- Work directly with international clients to clarify requirements, propose architectures and report progress

- Write clean, tested, well-documented Python code and review teammates' code

Requirements

- 5+ years of AI / Machine Learning engineering, with models shipped to production

- Strong Python and solid hands-on PyTorch

- Practical experience building LLM applications: prompt design, RAG, vector databases, agent frameworks

- Solid Computer Vision background (CNNs, Transformers, YOLO-family detectors)

- Model optimization and inference with ONNX and/or TensorRT

- Familiarity with MLOps tooling, Docker and at least one major cloud platform

- Good English to work directly with international clients

- Able to work independently and communicate clearly in a remote setup

Nice to have

- Deployment on NVIDIA Jetson (JetPack, DeepStream) or other edge accelerators

- Fine-tuning LLMs (LoRA/QLoRA) or serving them efficiently (vLLM, TensorRT-LLM, llama.cpp)

- Video analytics, multi-camera or real-time streaming pipelines

- C++ for performance-critical inference code

- Outsourcing, consulting or client-facing experience

Tech stack

Python, C++ (plus), PyTorch, Hugging Face, OpenCV, RAG, vector databases, agent frameworks, ONNX, TensorRT, NVIDIA Jetson, Docker, CI/CD, cloud platforms

What we offer

- 100% remote (candidates must be based in Vietnam)

- USD 2,000–3,000/month

- Freelance or second job welcome

- Flexible hours, with reasonable overlap with client time zones

- Work directly with international clients on real-world engineering projects