SF Tech Hiring Has Quietly Cracked Open Again, But the Jobs Look Nothing Like Before

The San Francisco tech job market added roughly 4,200 roles in the second quarter of 2026, according to data from the Bay Area Council Economic Institute released last week, but more than 60 percent of those openings carry an explicit AI-skills requirement that didn't exist in equivalent postings two years ago. The rebound is real. The catch is that it's restructuring who gets hired, how fast, and for how much.

This matters right now because the city is absorbing the shock of two overlapping forces at once. A wave of generative-AI tooling has compressed the time it takes a small engineering team to ship product, reducing headcount needs per startup even as venture capital flows back into the market. Crunchbase data shows SF-based startups closed $6.1 billion in funding during Q2 2026, up 34 percent year-over-year, yet total engineering headcount at seed and Series A companies is flat. More money, fewer seats.

SoMa Recruiters Are Rewriting Job Descriptions in Real Time

Walk into any WeWork on Howard Street on a Tuesday morning and you'll find recruiters doing something they weren't doing eighteen months ago: posting roles that blend traditional software engineering with what the industry is now calling "AI wrangling", prompt engineering, fine-tuning open-source models, and evaluation pipeline work. Hired.com, which operates its candidate marketplace out of offices near Caltrain at Fourth and King, says the median time-to-fill for a pure backend role in San Francisco has dropped from 47 days in January 2025 to 29 days today, mostly because companies have narrowed their searches and know precisely what they want.

Coding bootcamps are scrambling to keep up. App Academy, which has run cohorts out of its Montgomery Street location for years, overhauled 40 percent of its curriculum in March to center on AI-augmented development workflows. General Assembly's Market Street campus launched a six-week "AI Product Fundamentals" course in April; the $3,200 program sold out its first three cohorts within days of opening enrollment. Neither school will say publicly how placement rates have shifted, but both confirmed to The Daily SF that employer partnership requests are running ahead of 2024 levels.

The Salary Picture Is More Complicated Than Headlines Suggest

Entry-level salaries are under quiet pressure. For a junior software engineer role in San Francisco, the median base offer in June 2026 sat at $118,000, down from $134,000 in mid-2024, per Levels.fyi's most recent quarterly report. Senior engineers with demonstrated AI integration experience, however, are commanding $210,000 to $260,000 in base salary, and some roles at AI-native startups in the Dogpatch neighborhood, where a cluster of model-infrastructure companies has set up along Third Street, are attaching equity packages last seen during the 2021 peak.

The disconnect is generating real anxiety in neighborhoods like the Tenderloin and the Excelsior, where city-funded retraining programs have been trying to funnel residents into tech since the pandemic. The SF Office of Economic and Workforce Development's TechSF initiative, which placed 340 people into tech roles in 2024, is now recalibrating its curriculum for 2026 with a focus on AI operations and data labeling work, entry points that still exist even as pure coding jobs narrow.

For job seekers, the practical reality is blunt: technical skills alone are no longer sufficient. Hiring managers at several SoMa-based startups described, on background, a preference for candidates who can demonstrate they've shipped something using AI tooling in the last six months, a GitHub repo, a side project, anything with a commit history. Those without that kind of portfolio are finding the rebound largely invisible from where they're standing. The city's tech workforce is splitting into two tracks, and the gap between them is widening faster than most retraining programs can bridge.