ML Engineer Salary in CaliforniaBay Area, Los Angeles, San Diego, Sacramento. With state tax math.

California is not one labour market. Bay Area ML compensation is shaped by the concentration of T1 frontier AI labs and T2 hyperscalers, both of which lean heavily on equity grants tied to high-multiple valuations. Los Angeles supports a sizable autonomous-vehicle and entertainment-tech cluster but lacks the equity density of San Francisco. San Diego is dominated by a single large employer in mobile and AI silicon, plus a biotech-ML cluster around UCSD. Sacramento and the Central Valley are mostly state-government and agriculture-tech adjacent, with very few frontier-lab or hyperscaler outposts. The result is a four-tier intra-state distribution with a roughly 40 percent spread between the top and bottom metros for the same level.

Figure ca-1. California metro base-salary bands and L5 total compensation, synthesised from Levels.fyi region filters and BLS OEWS California 15-2051 (Data Scientists), as of May 2026. Bands are illustrative; individual offers vary substantially with employer tier and specialisation.

California has the highest top marginal state income tax in the United States. For an ML engineer earning a typical $250,000 base, the effective state income tax rate is approximately 8 to 9 percent. At total compensation levels above $1,000,000 (achievable at frontier labs or quant firms), the additional 1 percent Mental Health Services Tax applies, lifting the top marginal to 13.3 percent on income above $1,000,000. The interaction with federal AMT, ISO exercise treatment, and RSU vesting calendars makes equity-heavy California compensation materially harder to plan around than equivalent comp in no-state-tax states.

California (specifically the San Francisco Bay Area) is the only metro in the world where a senior ML engineer can routinely access total compensation packages above $500,000, and not by working harder or with more impressive credentials than peers elsewhere. The mechanism is equity at scale. Three concentric reasons explain why.

First, T1 frontier AI labs (foundation-model companies) are mostly headquartered in San Francisco proper or Mountain View. Their pre-IPO equity, granted on a 4-year vest, is currently valued on the basis of $30 billion to $300 billion paper valuations. At a $50 billion valuation with a 0.05 percent grant, that is $25,000,000 in paper value vesting over 4 years. Even a 90 percent valuation haircut at eventual IPO leaves $2.5 million. Senior IC offers from T1 labs in 2024 to 2026 have included grants in this range, with the trade-off that the entire upside is contingent on the lab continuing as a going concern.

Second, T2 hyperscalers (trillion-dollar public platform companies) headquartered in the Bay Area have run an arms race on RSU refresh grants since 2022 to retain ML talent that otherwise leaves for T1 labs. A senior L5 ML engineer with a 4-year initial RSU grant of $400,000 plus annual refresh of $100,000 to $200,000 is now common at the largest two or three hyperscalers. At public-company-stable share prices, this is realised cash, not paper.

Third, San Francisco quant trading firm outposts pay the largest cash bonuses in the city. A senior quant ML researcher at a systematic trading fund in San Francisco can earn $500,000 to $1,500,000 in performance-tied cash bonus on top of a $250,000 to $400,000 base. The bonus is cash and the partner track is a potential long-run upside, but there is no equity grant.

The cost of accessing this equity story is rent, taxes, and the implicit cost of living within commuting distance of San Francisco and Mountain View. Median two-bedroom rent in the Bay Area in 2026 remains above $4,500 per month, and California state income tax at $250,000 to $400,000 is meaningfully above the equivalent in Washington, Texas, or Florida.

The math: a Bay Area $300,000 base plus $150,000 equity (T2 hyperscaler L5) returns approximately the same take-home cash as a $260,000 base plus $130,000 equity in Seattle (T2 L5), once California state income tax and Washington's lack of state income tax are accounted for. But the Bay Area resident also has access to T1 frontier-lab equity that does not exist in Seattle, plus the option to switch employers without relocating. The structural premium is real; the after-tax delta is smaller than the headline numbers suggest.

Los Angeles supports a distinct ML labour market shaped by three industry concentrations: autonomous-vehicle and aerospace (with significant ML hiring at major OEMs and defence primes), entertainment technology (production-pipeline ML and content recommendation at streaming platforms), and a fast-growing AI-startup cluster centred on Pasadena and Santa Monica. LA L5 base ranges $180,000 to $210,000 with total compensation $210,000 to $300,000, roughly 12 to 18 percent below the Bay Area baseline at the same level. Equity grants are smaller and less frequently refreshed because the dominant employers are public companies with lower share-price growth than the Bay Area pre-IPO unicorns.

San Diego is dominated by Qualcomm (which has significantly expanded its AI organisation since 2023), biotech-ML around UCSD spinouts, and a defence-adjacent ML cluster (including ML-for-satellite work). San Diego L5 base falls in the $165,000 to $190,000 range with total compensation $185,000 to $260,000. The metro lacks T1 frontier-lab presence entirely. Compared to the Bay Area, San Diego trades roughly 25 to 35 percent lower nominal compensation for materially lower rent and a better climate; the after-tax purchasing-power gap closes to 15 to 20 percent.

Both LA and San Diego are net importers of ML talent who relocated from the Bay Area for lifestyle reasons. The trade-off is real: an LA L5 ML engineer at a major streaming platform earning $250,000 total compensation can afford a single-family home in a way that a Bay Area equivalent at $350,000 typically cannot. The closer the compensation is to base-only (rather than equity-heavy), the smaller the relocation penalty.

The Bureau of Labor Statistics Occupational Employment and Wage Statistics (OEWS) programme publishes California-specific wage data annually for SOC codes 15-2051 (Data Scientists) and 15-1252 (Software Developers). ML engineer is not yet a separate SOC code, so the best public sources are these adjacent categories plus Levels.fyi self-reported data filtered by California region. The BLS California state-level page shows 15-2051 median annual wage above $150,000 (2024 vintage) and 90th percentile above $230,000. ML engineer compensation runs higher than the broader Data Scientist category because the title is more recently established, the labour supply is tighter, and equity packaging is more common.

Within California, the San Francisco / Oakland / Fremont MSA and the San Jose / Sunnyvale / Santa Clara MSA both rank in the top three metros nationally for 15-2051 annual mean wage. The Los Angeles / Long Beach / Anaheim MSA ranks roughly 10th to 15th. The Riverside / San Bernardino / Ontario MSA, despite high population, has lower wages and almost no T1 / T2 employer presence.

A caveat: BLS OEWS captures W2 wage data, which usually excludes equity (RSU vesting is reported as wage in the year of vesting, so the data does include some equity-driven inflation but lagged). The true total-compensation distribution in the Bay Area is meaningfully wider in the upper tail than OEWS suggests, because equity-vesting events are concentrated in a small number of high earners. Levels.fyi and Blind self-report data captures this tail better, but with self-selection bias toward higher-compensated respondents.

1. BLS Occupational Employment and Wage Statistics, California, 2024
2. Levels.fyi Software Engineer, San Francisco Bay Area
3. California Franchise Tax Board 2026 personal income tax tables
4. California Employment Development Department wage and employment statistics
5. BEA Regional Price Parities by metro