You've seen the salary surveys. "$110K–$180K for a Senior Software Engineer." Technically true. Practically useless. A Senior SWE at a 40-person Series B in Columbus and a Senior SWE (L5) at Google in Seattle are doing nominally comparable work — the difference in total compensation can be $150,000 a year. That's not a range. That's a different life.
This report cuts the data by what actually matters: your specialization, your level, and the tier of company you're at. It covers software engineering, AI/ML, product management, data science and analytics, and security/infrastructure. Where relevant, it includes total compensation — base plus equity plus bonus — because that's the number that actually goes to your bank account.
All base figures are in USD for major US markets (NYC, SF, Seattle, Chicago, Austin). For remote roles, employer-tier benchmarks apply more than geography. International context is flagged where the gap is meaningful.
How to read the tables
Ranges show the 25th–75th percentile base salary for that role/level in major US markets. The "top-of-band" figure is a real data point, not a ceiling — 25% of people at that level earn more. Equity and bonus figures are annual equivalents and addressed separately. Remote roles skew toward the employer's HQ tier, adjusted down ~10–20% for non-coastal locations.
The Framework
How Tech Compensation Actually Works
Before any number makes sense, you need the mental model. Tech compensation at most companies above ~100 employees has three components that matter:
Total Compensation = Base + Equity (annualized) + Bonus
At hyperscalers (Google, Meta, Microsoft, Amazon), equity can exceed base. At startups, base is higher relative to equity, which is in illiquid options.
The "employer tier" framework is the single most important lens for reading tech comp data. It's not about prestige — it's about how each tier prices talent and structures offers.
Throughout this report, salary ranges are anchored to Tier 2 as the baseline — it represents the broadest and most accessible market for most tech professionals. Tier 1 premiums and Tier 3 discounts are noted where the gap is significant.
01 / Software Engineering
Software Engineering Salaries
The biggest function, the widest range, and the level that matters more than the title.
Leveling systems differ wildly across companies, but a rough universal translation holds: Junior (0–2 yrs), Mid (2–5 yrs), Senior (5–8 yrs), Staff (8–12 yrs, or equivalent IC leadership), Principal / Distinguished (above that). The important caveat: years of experience are a weak proxy. A highly capable engineer can reach Senior in 3 years; a coasting engineer might never leave Mid.
Use the scope definition, not the year count, to calibrate your level before benchmarking.
| Level | Scope Signal | Base (Tier 2) | Est. TC (Tier 1) | Base (Tier 4) |
|---|---|---|---|---|
| Junior / L3 | Guided tasks, defined scope | $90K–$125K | $180K–$230K TC | $75K–$100K |
| Mid / L4 | Independent features, some mentoring | $120K–$160K | $230K–$300K TC | $95K–$130K |
| Senior / L5 | Owns systems, drives tech decisions | $155K–$210K | $300K–$420K TC | $120K–$160K |
| Staff / L6 | Cross-team impact, architecture ownership | $195K–$260K | $420K–$580K TC | $145K–$190K |
| Principal / L7+ | Org-wide or company-wide technical direction | $240K–$330K | $550K–$900K+ TC | $170K–$230K |
TC = Total Compensation (base + annualized RSU + bonus). Tier 1 TC figures sourced from Levels.fyi aggregates, Q4 2025.
The IC vs. Management Fork
At most large tech companies, there are two valid career tracks past Senior: the Individual Contributor (IC) track (Staff → Principal → Distinguished → Fellow) and the Engineering Manager track. Compensation is roughly equivalent at each level equivalency — a Staff SWE and an Engineering Manager leading 6–8 people earn similar base in the same company.
The choice between them is about what work you want to do, not which one pays more. What it's not: a seniority ladder where you have to manage people to advance. The IC track at hyperscalers goes all the way to Fellow-level compensation in the multiple millions. Don't take the management track because you think it's the only way up.
Specialization Premiums on Top of SWE Base
Not all Senior SWE roles are priced equally. These specializations consistently earn above the generalist SWE band:
02 / AI & Machine Learning
AI / ML Engineering Salaries
The highest-demand discipline in tech. The premium is structural, not temporary.
The AI/ML engineering market in 2026 is split into two distinct segments: applied ML engineers who build and deploy models within products (the larger group), and frontier ML researchers who advance the underlying science (the smaller, extraordinarily well-compensated group). If you're not at a lab, you're almost certainly in the applied segment — and that's where most of the hiring is happening.
Every major enterprise is deploying AI into products, workflows, and infrastructure. The bottleneck is people who can actually build production-ready AI systems — not just train a model in a notebook, but deploy, monitor, and iterate on it at scale. That scarcity is reflected in comp.
| Role | Level | Base Range | Est. TC (Tier 1) |
|---|---|---|---|
| ML Engineer | Mid | $145K–$190K | $270K–$360K TC |
| ML Engineer | Senior | $185K–$255K | $360K–$520K TC |
| ML Engineer / Staff | Staff | $230K–$310K | $500K–$750K TC |
| AI Research Scientist | Mid–Senior (Lab) | $220K–$340K | $450K–$900K+ TC |
| LLM / GenAI Engineer | Mid–Senior | $170K–$260K | $320K–$500K TC |
| MLOps / AI Infra Engineer | Senior | $165K–$230K | $310K–$450K TC |
Research Scientist figures reflect top-tier AI labs (Anthropic, OpenAI, Google DeepMind, Meta AI). Applied ML at non-lab companies is the row above.
If you're adjacent to AI but don't have the title yet:
Data engineers who've built LLM pipelines, backend engineers who've integrated AI APIs into production systems, and analytics engineers who've built RAG workflows are commanding ML Engineer-adjacent comp. The credential matters less than the demonstrated production experience. Document what you've shipped.
03 / Product Management
Product Manager Salaries
High comp ceiling, wide variance, and a title that means wildly different things across companies.
Product Management is arguably the highest-variance role in tech when it comes to comp. At a hyperscaler, a Senior PM runs a product area touching millions of users, earns $250–350K+ total comp, and has the organizational leverage to make decisions worth hundreds of millions. At a 30-person startup, a "Senior PM" might be the only PM and earn $130K with minimal equity. The title is almost irrelevant. Scope is everything.
| Level | Scope Signal | Base (Tier 2) | Est. TC (Tier 1) |
|---|---|---|---|
| Associate PM | Feature-level, high mentorship | $95K–$130K | $190K–$240K TC |
| PM | Product area ownership, independent | $130K–$175K | $240K–$330K TC |
| Senior PM | Multi-team, revenue or retention impact | $155K–$215K | $300K–$430K TC |
| Principal / Group PM | Multiple PMs, P&L ownership | $195K–$265K | $400K–$600K TC |
| Director / VP of Product | Business unit, full product strategy | $230K–$340K | $500K–$900K+ TC |
The AI PM premium is real
PMs with genuine technical depth in AI/ML — who can work credibly alongside research and engineering teams on model-adjacent products — are seeing 20–30% comp premiums over comparably leveled generalist PMs. If your current product involves any AI surface area and you're building that fluency, it's worth calling out explicitly in any negotiation.
04 / Data
Data Science & Analytics Salaries
Three distinct tracks with different comp ceilings. Know which one you're actually on.
"Data" is not one career. The field has matured into three distinct tracks, each with different work, different skills, and different comp trajectories. Where you sit matters for benchmarking.
| Role | Level | Base Range | Note |
|---|---|---|---|
| Data Scientist | Mid | $115K–$155K | Varies heavily by industry vertical |
| Senior Data Scientist | Senior | $145K–$200K | Bonus 10–20%; some RSU at scale cos |
| Staff Data Scientist | Staff | $185K–$250K | Effectively a science leadership role |
| Data Engineer | Mid–Senior | $130K–$195K | Strongest demand growth in data |
| Analytics Engineer | Mid–Senior | $115K–$170K | Fast-growing role; newer title |
| Data Analyst | Mid–Senior | $85K–$130K | Ceiling lower without engineering skills |
The single most reliable way to increase your comp in data is to push up the engineering skill stack. A data scientist who can build production pipelines, or an analyst who can write dbt models and own a data product, earns in a different bracket. It's not glamorous career advice — it's just where the market prices the work.
05 / Security & Infrastructure
Security & Infrastructure Salaries
The most consistently underestimated comp segment in all of tech. If you're here, you're likely underpaid.
Security engineering is structurally underpaid relative to its market value and its criticality. The reason is cultural: security teams are often staffed as cost centers, not value generators. Boards understand this after a breach. The comp market knows it before one happens — which is why experienced security engineers at companies that do understand the stakes earn more than their headcount would suggest.
Platform / infrastructure engineering is in a similar position: less visible than product engineering, but often the highest-leverage work in the stack. Senior infra engineers who have rebuilt observability systems, tamed cloud costs, or built developer platforms are increasingly in demand and increasingly aware of it.
| Role | Level | Base Range | Context |
|---|---|---|---|
| Security Engineer | Mid | $130K–$175K | Appsec, cloud sec, detection |
| Security Engineer | Senior | $175K–$240K | +25–30% vs. generalist SWE same level |
| Security Architect | Sr. / Staff equiv. | $195K–$270K | Fintech/healthcare pay top of band |
| Platform / Infra Engineer | Senior | $165K–$225K | Cloud cost optimization a hot skill in '26 |
| DevOps / SRE | Senior | $155K–$215K | SRE (Google model) skews higher |
| CISO | Executive | $260K–$450K+ | Significant equity; highly variable |
06 / Equity
Equity: What You Actually Need to Know
RSUs are real money. Options are a bet. Neither is complicated once you understand the mechanics.
RSUs (Public Companies)
- Vest over 4 years, typically quarterly after year 1 cliff
- Taxed as ordinary income when they vest
- Price at vesting = your actual comp
- Refreshes: high performers get new grants annually
- Rule of thumb: A $200K RSU grant = $50K/year in cash, pre-tax
Options (Private Companies)
- Right to buy shares at the strike price (usually set at last 409A valuation)
- Only valuable if the company exits above your strike price
- ISOs vs. NSOs have different tax treatment — read the difference
- You often have 90 days post-departure to exercise (or lose them)
- Rule of thumb: Worth $0 until there's a liquidity event
The honest question to ask before accepting options:
"If this company exits in 5 years at 3× the current valuation, and I exercise at my strike price, what do I actually take home after dilution and taxes?" If you can't answer that, the number in your offer letter is theoretical. Ask the recruiter for the current cap table details, preferred shares outstanding, and last 409A price. Good companies give you this. The answer tells you a lot.
07 / Negotiation
Using This Data in a Negotiation
Benchmarks only matter if you can put them on the table. Here's how.
The data in this report is only useful if you know how to deploy it. Walking into a negotiation and saying "the internet says this role pays $180K" is not a strategy. Here's what actually works.
For a full negotiation playbook — scripts, tactics for stalled offers, and handling exploding deadlines — see the Salary Negotiation Playbook (C-1-P) and Exactly What to Say in a Salary Negotiation (C-1-C2).
FAQ
Frequently Asked Questions
Key Takeaways
What to do with this data
Benchmark on TC, not base. At Tier 1 companies, base can be less than half the story. Never compare offers on base alone.
Calibrate your level by scope, not title. Your employer's leveling system is not the market's. Find your equivalent using scope signals.
AI/ML and security premiums are real and structural. If you have these skills, price them explicitly — don't wait for a manager to notice.
The IC track and the management track pay the same. Choose based on the work you want to do, not which one you think pays more.
Options are a bet. RSUs are income. Don't let a large option grant replace a competitive base at a stage where the outcome is genuinely uncertain.