Junior AI Engineer - GenAI Applications
Company: KY LONG SERVICE TRADING INVESTMENT COMPANY LIMITED
Website: [protected info]
Team: Data & AI Product Development
Location: 12th Floor, 27 Dinh Bo Linh Street, Binh Thanh Ward,Ho Chi Minh City, Vietnam
Employment type: Full-time
Salary range: Negotiable
About the role
We're hiring a Junior AI Engineer to help build LLM-powered features - retrieval-augmented generation (RAG), agentic workflows, and prompt-driven product capabilities - into KY LONG CO., LTD. products. You'll work alongside senior engineers turning model APIs into reliable, measurable features that real users depend on.
This is a hands-on, ship-oriented role. You won't be training foundation models; you'll be building the application layer on top of them: pipelines, retrieval, evaluation, and the glue that makes GenAI actually work in production.
What you'll do
- Build and iterate on RAG pipelines: chunking, embedding, retrieval, and grounded prompt construction.
- Integrate LLM provider APIs into product features with sensible error handling, retries, and cost/latency awareness.
- Write and refine prompts, then measure their quality with evals rather than vibes.
- Develop small agent/tool-use workflows and the supporting plumbing (function calling, structured outputs).
- Write clean, tested Python; participate in code review; document what you build.
- Debug behavior of non-deterministic systems and contribute to observability (logging, tracing, eval dashboards).
What we're looking for (must-have)
- Solid programming fundamentals, ideally in Python - comfortable with data structures, functions, and writing readable code.
- Working understanding of computational complexity - can reason about the time/space cost of code in Big-O terms, recognize when an approach won't scale, and pick a more efficient data structure or algorithm.
- Working understanding of how LLMs are used in applications: prompting, context windows, embeddings, and RAG at a conceptual level.
- Fundamental grasp of machine learning - supervised vs. unsupervised learning, overfitting/underfitting, the train/validation/test split, and how to read evaluation metrics (accuracy, precision/recall, F1).
- Fundamental grasp of statistics - distributions, mean vs. median, correlation vs. causation, basic probability, and how to interpret a result (e.g. what a p-value does and doesn't say).
- Minimal hands-on experience with core data science / ML toolkits - comfortable with NumPy and polars for working with data, and at least passing familiarity with a deep-learning framework (PyTorch or TensorFlow) and HuggingFace Transformers. Depth isn't expected for a junior; having actually used a few of these is.
- Ability to read API docs and integrate a third-party service end to end.
- Clear communication and a habit of testing your own work.
- 0-2 years experience; new grads, bootcamp grads, and strong self-taught candidates are also welcome.
Nice to have
- Experience storing and querying embeddings using a database's built-in vector support (e.g. Postgres/pgvector, Elasticsearch).
- Familiarity with an agent/orchestration framework - LangGraph preferred (which implies comfort with the underlying LangChain ecosystem), - or strong, well-reasoned opinions about doing without one.
- Exposure to evaluation tooling, prompt versioning, or LLM observability.
- Git, basic cloud (e.g. AWS, GCP, Azure), containers, or CI experience.
- A side project, repo, or demo that shows you've actually built something with LLMs.
What we offer
- Mentorship from senior AI engineers and real production ownership early
- [Learning budget, conference/credits, compute]
- [Health insurance, leave, other benefits]
- [Equipment]