1. 開發聊天系統語音合成、開發歌聲合成系統。
2. 熟悉現在熱門資料庫、爬蟲、資料搜尋。
3. 熟悉TRANSFORMER、PROMPTING、PRETRAINING (預訓練不是微調!)、INSTRUCT TUNING。
4. 熟悉Python、Pytorch、Keras,機器學習(machine learning)的各類算法,並有實際項目使用經驗。
5. 熟悉其中一門架構Bert、Attention、Transformer、decoder ONLY、Vocoder、Unsupervised Learning、K means比較。
6. 熟悉資料處理比重調整、跑過模型預訓練、TOKEN 性質調整、STREAMIN、DUMMY、QUANTIZATION、CONTEXT WINDOW、調整LLM SERVING、向量相識度搜尋、RAG資料解鎖、Self-RAG、ATTENTION 調整: PAGED ATTENTION、WINDOW ATTENTION。
+++++主動積極,提出解決方案和改進建議,我們喜歡這樣的團隊成員!+++++
1. Develop chat system voice synthesis, and singing voice synthesis systems.
2. Familiar with popular databases, web scraping, and data searching.
3. Proficient in TRANSFORMER, PROMPTING, PRETRAINING (pretraining, not fine-tuning!), and INSTRUCT TUNING.
4. Skilled in Python, Pytorch, Keras, and various machine learning algorithms with practical project experience.
5. Proficient in one of the architectures such as Bert, Attention, Transformer, decoder ONLY, Vocoder, Unsupervised Learning, and K-means comparison.
6. Experienced in data processing weight adjustment, model pretraining, TOKEN attribute adjustment, STREAMIN, DUMMY, QUANTIZATION, CONTEXT WINDOW, LLM SERVING adjustment, vector similarity search, RAG data unlocking, Self-RAG, ATTENTION adjustments: PAGED ATTENTION, WINDOW ATTENTION.
+++++Be proactive; we love people who are proactive and suggest solutions and improvements in the team's work!+++++
1. 有搭建雲端的機器學習系統 (cloud-based machinelearning、TensorRT、Triton Inference Server)者優先。
2.公司使用A100、H100大量訓練。
**應徵者需接受公司測試**