Applied Machine Learning Engineer
Engineering5+ yearsRemote · Bangalore, India
Customers arrive at Vidman AI to fine-tune open models on dedicated GPUs, metered by the second, and to serve what they train through one API. As an Applied Machine Learning Engineer you make their models work: you take fine-tuning jobs from dataset to deployed endpoint, build the tooling that makes that path boring, and turn individual customer problems into platform capabilities.
What you'll do
- 01Own fine-tuning engagements end to end: dataset inspection, method selection, run configuration, evaluation, and handoff.
- 02Build internal tooling for dataset validation, training orchestration, and model evaluation across a catalog of 250k+ open models.
- 03Debug the strange ones: loss spikes, adapter pathologies, tokenisation edge cases, VRAM ceilings.
- 04Feed recurring customer patterns back to product and research as concrete, prioritised proposals.
- 05Help customers serve their fine-tuned weights through the inference stack.
What you bring
- At least 5 years applying machine learning in production environments.
- Real experience fine-tuning LLMs or vision-language models — configuring runs, reading curves, not only calling APIs.
- Strong Python and PyTorch, comfortable reasoning about GPU memory when it gets tight.
- Data instincts: the dataset gets looked at before the model.
- Clear communication — a failed run explained to a customer without hand-waving.
Nice to have
- Breadth across model families — DeepSeek, Qwen, Kimi, GLM, Gemma, and the rest.
- Depth in the Hugging Face ecosystem: transformers, PEFT, TRL, datasets.
- Multi-GPU training frameworks and checkpoint management, in anger.
- Serving experience — vLLM or a similar inference stack.