Machine Learning Engineer

About the Project

The project develops a distributed platform that leverages decentralized GPU power for executing meaningful AI computation tasks.

You’ll join the R&D team responsible for designing, testing, and optimizing distributed inference workflows.

Responsibilities

• Conduct inference research and benchmarking on models like DeepLabV3, YOLO, BERT, CLIP, Wav2Vec2.




• Implement slicing, and merging scripts for model evaluation.




• Implement and validate model-splitting and distributed inference strategies.




• Support fine-tuning and model adaptation for selected inference workloads.




• Collaborate with the Task Manager backend team to define interfaces, task schemas, and data contracts for distributed workloads.




• Prepare R&D documentation: experiment summaries, reports, and optimization recommendations.




• Participate in regular R&D meetings and contribute to component-level design discussions.

Requirements

• 3+ years of experience as a Machine Learning Engineer focused on model inference systems or optimization/fine-tuning




• Solid hands-on experience with PyTorch, TorchVision, and Hugging Face Transformers.




• Proven ability to analyze trade-offs between accuracy, performance, and hardware constraints.




• English – upper-intermediate or higher (for documentation and team communication).

Nice to Have

• Understanding of GPU memory management, latency profiling, and multi-GPU environments.




• Familiarity with distributed computation frameworks (Ray, Dask, or custom message-based orchestration).




• Previous work on AI compute marketplaces, federated learning, or distributed AI inference.