The median starting salary is $200,000 a year, measured on the 152 postings that publish an annual salary range in US dollars. Measured September 2026. Compare with AI salaries by field.
Fine-tuning jobs in San Francisco: 195 open roles (Sep 2026)
Live postings that name fine-tuning in the job title or description, located in San Francisco, United States. Each one is rated AI Level 1–4 for how much AI is in the daily work.
- open roles
- 195
- AI Level 3+ (AI in the daily work)
- 165
- remote
- 43
- median starting salary
- $200,000
AI level of Fine-tuning jobs in San Francisco
- AI Level 4126 (65%)
- AI Level 339 (20%)
- AI Level 20 (0%)
- AI Level 130 (15%)
Companies hiring for Fine-tuning in San Francisco
Fine-tuning jobs in other cities
Roles that ask for it
Common questions
How much do Fine-tuning jobs in San Francisco pay?
How much AI work do Fine-tuning jobs in San Francisco involve?
Of the 195 live postings: AI Level 4: 126 (65%); AI Level 3: 39 (20%); AI Level 1: 30 (15%). See what each level means. Counted September 2026.
Which companies are hiring for Fine-tuning jobs in San Francisco?
By number of live postings: Together AI (55), Anthropic (17), Scale AI (15), OpenAI (8), Waymo (8). Counted September 2026.
Which other cities have Fine-tuning jobs?
Cities with their own list, by number of live postings: New York (102), Seattle (69), Santa Clara (48), San Jose (40), Mountain View (29). Counted September 2026.
Which skills come up most in Fine-tuning jobs in San Francisco?
Named in the title or description, with the number of postings that mention each: LLM (108), Reinforcement learning (106), Python (103), Machine learning (88), PyTorch (64). Counted September 2026.
Which roles are open in Fine-tuning jobs in San Francisco?
By number of live postings: Machine learning engineer (25), AI engineer (10), Research scientist (10), Solutions architect (7), AI product manager (5). Counted September 2026.
A posting counts when it names Fine-tuning in its title or description. Counted September 2026 on live postings; the list refreshes every few hours. Levels follow our method.