Job description for AI Engineer (Working in Da Nang) at SotaTek
JOB DESCRIPTION:
RAG Development:
- Design, build, and optimize end-to-end RAG pipelines (ingestion → indexing → retrieval → reranking → generation).
- Integrate and tune vector databases (Pinecone, Weaviate, Milvus, Chroma, etc.).
- Improve chunking strategies, embedding quality, and similarity search performance.
- Build evaluation pipelines for RAG (precision/recall, context relevance, response quality).
Agentic AI / AI Agents:
- Develop AI Agents using frameworks like LangGraph, AutoGen, or LlamaIndex Agents.
- Design multi-agent workflows (planner → executor → evaluator).
- Implement tool/function calling, serverless integrations, and dynamic reasoning workflows.
Computer Vision & Multimodal AI:
- Support the development of multimodal AI applications by integrating Large Language Models (LLMs) with Computer Vision models.
- Contribute to Computer Vision use cases such as Image Segmentation, image understanding, and intelligent document processing where applicable.
Model Context Protocol (MCP):
- Design and implement MCP servers exposing tools, APIs, and datasets to LLMs.
- Integrate MCP into application stack (frontend ↔ backend ↔ LLM).
- Build custom MCP tools (database connectors, internal API tools, scraping tools, etc.).
System Architecture & MLOps:
- Build and maintain inference infrastructure (Docker, Kubernetes, GPU servers).
- Implement logging, monitoring, and observability for RAG/Agent pipelines.
- Optimize model serving cost and performance (quantization, vLLM, batching, TensorRT, etc.).
Requirements:
- Technical foundation: 2-4 years experience with AI Engineer
- AI knowledge: Computer Vision, Speech AI (STT, TTS, Voice AI), Generative AI (Image, Video), Predictive Machine Learning, MLOps / LLMOps
- Data skills: Experience with data preprocessing, feature engineering, and model evaluation
- Experience with building RESTful APIs and working with SQL/NoSQL databases
- Problem-solving aptitude: Ability to approach complex problems methodically and develop creative solutions
- Continuous learning mindset: Enthusiasm for staying current with AI research and implementing new techniques
- Version control: Proficiency with Git and collaborative development workflows
- Fluent in English or Japanese is a plus
Compensation & Benefits:
🎯Flexible working regime and health care:
- Flexible timekeeping (from 8:00 - 9:00 to 17:30 - 18:30)
- Minimum 14 paid leaves per annum for all employees after probation
- 01-day remote work per month
- A flexitime allowance of 90-180 minutes per month for employees
- 01 hour paid leave per day for women having children under 12 months
- Social insurance, health insurance, unemployment insurance and Bao Viet care insurance
🎯Transparent and fair benefits:
- Saturday & Sunday OFF, Overtime pay is 150%, 200%, 300% as per labor law;
- Work performance review 2 times/ year (in April and October)
- 13th-month salary
- Bonus Policy: Public holidays (2/9, 30/4, 1/5, 1/1,...); Personal Performances; Excellent Team; Performance bonus in Token of the project;..
- Men’s Day, Women’s Day, Children’s Day, Mid-Autumn Festival and other benefits under the provisions of the company
🎯Dynamic environment and open culture:
- Year-end party, sports day, yearly company trip and quarterly team building,...with a generous budget
- Socialize with colleagues through monthly Happy Hour
- Monthly allowance when joining clubs: Soccer, Swimming, Yoga, Music,...
- Nice & modern working space with young, dynamic & friendly colleagues and free coffee, tea, drinks,...
- Flat, open and sharing culture with friendly management team; outsourcing company with product mindset
🎯Strong learning culture:
- Free training courses for technical and soft skills (presentation skills, communication skills, foreign language courses,...)
- Account to log in to our online learning system, which contains thousands of valuable lectures (LMS)
- Participate in workshops, seminars, tech talk,... with sharing from experts inside and outside the company
- Working opportunities with technical gurus who built and operated world-class applications with millions of users.
