Pros
- Overseas exposure in one of the most dynamic (and polluted) cities in the world - Beijing - Work directly with medical professionals - Opportunities to have hands-on experience in AI model building and using deep learning frameworks - The medical datasets available for model building have unprecedented size - Opportunity to work on a multiple servers with multiple Nvidia top-of-the-line GPUs - Great colleagues who can do good work (AI, ML, SE)
Cons
- Overseas deployment terms were unclear at the start of the interview process. - Poor or non-existent tech infrastructure (data pipelining, ETL, CI/CD), messy data management, lack of data security / backup infrastructure - Management is poor and does not take feedback nor improve on past mistakes. The entire company is controlled by the handful of Chinese investors. Timelines are set by those businessmen who have no real technical knowledge - they only know how to demand results without much awareness of technical feasibility. AI problems don't work by simply throwing more $ at it. This rushed timeline also resulted in a hasty application process for certifications such as ISO and CE - Rushed timelines due to the above reason, leading to alot of technical debt. Deep and fundamental research problems which other models depend on first being solved, cannot be tackled properly, resulting in many models that are built in a rush but which cannot be used in the long-term, or are unfeasible for real clinical use in the long run