【What to expect】 Here at iKala, we are developing new algorithm and data pipeline which powers business decision in marketing and in-house analytic platform.
You will cowork with teams of working in Machine Learning, Data Acquisition, Data Scientist and Data Analysis to develop or optimize algorithm on large scale multi-modal data.
【What you will be doing】
Develop in house multilingual NLP models ( Chinese, English, Japanese ) Tuning and optimize algorithm to handle large scale data ( > 100 GB size ) Carry out ablation studies and experiments. Read latest papers and implement state of the art model. 【Requirements】
Master Degree or above in engineering or quantitative discipline ( Statistic, Mathematics, Computer Science, Electrical Electronic ) Good communication and writing skills. Experience in common machine learning frameworks ( scikit-learn, transformers, onnx, PyTorch) Open to learn new languages (golang, C) and frameworks. Comfortable working with Git. Full understanding in Transformers and related research ( sparsity, LLM ) 【Pluses】
Published papers in reputable international conferences ( KDD, ACL, WWW, CVPR, ECCV, ICLR, AAAI, NeurIPS ) Past experiences in contributing to open source projects. Experience in GPU speedup tricks ( mix precision, distributed training ) Worked in Transformer, CLIP or self supervised learning related research or side projects. Desire to work in fast pace startup environment. more info»>https://ikala.tv