Skill diminta
DockerEnglishFastAPIPyTorchPython
Deskripsi
- Deskripsi pekerjaan Machine Learning Engineer (RnD) PT Karsa Omni Teknologi Adicipta
- we're looking for a Machine Learning Engineer (R&D) who thrives in the space between experimentation and execution.
You don't need a PhD or a publication list — but you do need curiosity: the kind that makes you fine-tune a model just to see what happens, or prototype a RAG pipeline over a weekend. You've worked on real AI projects in industry, shipped code that touched users, and know that the best research is the kind that survives contact with production.
If you're mid-level with strong potential, or experienced but still hungry to explore — and you care more about what works than what's novel — this is your kind of challenge.
This is a hybrid role based in West Jakarta, where curiosity meets code.
What You'll Do
- ✅ Prototype and validate AI/ML concepts for real-world use cases: NLP, RAG, agents, or GenAI workflows
- ✅ Build lightweight pipelines for data prep, model training, evaluation, and serving
- ✅ Experiment with LLM tooling: prompt engineering, fine-tuning, retrieval, caching, fallback logic
- ✅ Collaborate with product and engineering to turn promising ideas into testable features
- ✅ Document experiments, measure impact, and iterate based on results — not just intuition
- ✅ Stay sharp on emerging techniques, but filter them through a lens of practicality
Who You Are
- ✅ 2–4 years in ML engineering, applied AI, or related roles — industry experience preferred, but strong mid-level candidates welcome
- ✅ Strong in Python and modern ML frameworks (PyTorch, Hugging Face, LangChain, LlamaIndex)
- ✅ Hands-on with LLM workflows: RAG, prompt design, evaluation, basic fine-tuning
- ✅ Comfortable with vector databases (Pinecone, Weaviate, pgvector) and API integration
- ✅ Bonus: Experience with MLOps basics (MLflow, Docker, FastAPI) or workflow tools (Dify, n8n)
- ✅ Fluent in English — written and spoken
- ✅ Nerdy enough to read a paper for fun — pragmatic enough to ask "does this actually help?"