Job Description

Design and implement novel research ideas. Ship state of the art models to production. Deep connections to academia. We do not delineate strongly between engineering and research. Everyone will contribute to writing production code and conducting research depending on individual interest and organizational need. Working on Machine Learning, you will: Design, train and improve upon cutting-edge models Develop new techniques to train and serve models safer, better, and faster Train large scale models on massive datasets Engage in research collaborations with our partner organizations and academic affiliations Publish your work in top-tier conferences and journals Work closely with product teams to develop solutions Assist our legal teams with preparation of patents on developed IP Join us at a pivotal moment, shape what we build and wear multiple hats! You may be a good fit if you have: Proficiency in Python and related ML frameworks such as PyTorch, Tensorflow for model prototyping and development Proficiency with optimized serving frameworks such as TF-Serving, Triton, Pytorch Serving Model optimization frameworks such as TensorRT, JAX, DeepSpeed, FasterTransformers, and XLA/MLIR Experience and proficiency in model quantization and knowledge distillation techniques Experience in building and scaling large-scale distributed training with multi-GPU/multi-Node set up Familiarity with modern deep learning architectures such as Transformers and their use in autoregressive/generative use cases Strong communication and problem-solving skills A demonstrated passion for applied NLP models and products Bonus: experience writing kernels for GPUs using CUDA Bonus: experience training on TPUs Bonus: paper at top-tier venues (such as NeurIPS, ICML, ICLR, AIStats, MLSys, JMLR, AAAI, Nature, COLING, ACL, EMNLP) This is neither an exhaustive nor necessary set of attributes. Even if none of these apply to you, but you believe you will contribute to Symbl, please reach out. We have a wide variety of backgrounds at Symbl.

foundit