On this page
💻 Tech & Engineering
Software Engineering
Build digital products. One of the highest ceilings in any field.
Entry earning
$110k
Junior Engineer
Senior tier earning
$400k
Staff / Principal
Source: Industry compensation surveys and public offer data (BLS does not track these roles separately)
Words with a dotted underline are industry terms. Hover or tap for a plain-English definition.
What the job actually looks like
Software engineers write the code that makes apps, websites, and systems work. A typical day is a mix of writing code, fixing bugs, reading other people's code, and meetings about what to build next.
Paths inside the field
Pick a branch to retrace the arc
Click any specialization. The timeline and salary curve above update to match.
The full arc
How this career unfolds
Junior Engineer
Total comp including stock
Mid-level Engineer
Senior Engineer
Staff / Principal
Top L6+: $500k to $900k
Surprising fact
A staff engineer at a top AI lab now out-earns most investment-banking MDs, cardiothoracic surgeons, and partners, often before age 30.
Lifestyle & optionality
The day-to-day reality
Typical hours
40 to 50/wk
High
Very high
Very high
Very high
Location is a gate
Top pay still favors Bay Area and Seattle, but remote work made this the most location-flexible high-paying career in America.
Fully remote roles are common. Skills transfer across industries.
Where this is heading
10-year outlook
Projected job growth (2023 to 2033)
Source: U.S. Bureau of Labor Statistics
AI and automation risk
The job will exist but look fundamentally different in 10 years.
Tailwinds
- AI is multiplying productivity, not replacing senior engineers
- ML/AI specialization paying record comp
Headwinds
- Entry-level jobs being absorbed by AI tooling
- FAANG hiring down 30-40% from 2021 peak
Real wages vs. inflation
Real wages at senior levels up, but junior/mid compression is real. Entry-level offers down ~15%.
Inside the pipeline
The gates, the timing, and what happens after
The narrow
The narrow gate
What it takes to land at or a top AI lab
Who actually gets there
new-grad acceptance rates run ~2 to 5% of applicants. Top AI labs (OpenAI, Anthropic, DeepMind) hire in the low hundreds per year from tens of thousands of applicants — under 1%. Reaching Staff / Principal engineer (the $500k+ tier) is roughly 1 in 20 of those hired.
What you actually need
- A strong CS program helps but isn't required. for recruiting: Stanford, MIT, CMU, Berkeley, UIUC, Georgia Tech, Michigan, Waterloo, UT Austin, UW. Outside these, you compete on projects and LeetCode.
- For AI labs (OpenAI, Anthropic, Google DeepMind): a from a top-20 ML program OR a real publication at NeurIPS / ICML / ICLR is effectively the floor for research roles.
- GitHub with real projects, not tutorial clones. At minimum one project people actually use or star.
- LeetCode: comfortable with ~200 medium / 50 hard problems, plus system design for senior roles. Interviews are algorithmic, not 'tell me about a project'.
- Internship-to-offer pipeline is huge. converts 60 to 80% of interns to full-time; getting the sophomore/junior internship is the main .
- Total comp at top tech is 60 to 70% stock. A $300k offer is roughly $180k base + $120k/yr in . Know the difference when comparing offers.
When to do what
When to do what (as of 2026)
The / top AI lab timeline as of 2026
Tech recruiting has moved meaningfully earlier - sophomore-summer programs are now the main junior-SA feeder. AI lab recruiting rewards published research or obviously impressive side projects.
HS senior year
Attend a strong CS program: Stanford, MIT, CMU, Berkeley, UIUC, Georgia Tech, Michigan, Waterloo, UT Austin, UW. Outside this list, project portfolio becomes the differentiator.
College freshman fall
Start LeetCoding. Goal: 100 mediums by end of freshman year. Ship one real side project to GitHub.
Freshman summer
Any real SWE internship (startup, regional tech, research lab). Meta / Google / Microsoft have freshman-specific programs - apply Oct-Dec of freshman fall.
Sophomore fall (Sept-Nov)
sophomore-summer programs open. Google STEP, Meta University, Microsoft Explore. Apply the week apps open.
Sophomore summer
Sophomore program OR strong-brand SWE internship. This is now the main feeder to junior-year FAANG SA.
Junior fall (Aug-Oct)
Junior summer / top-tech recruiting. Apply to 15 to 25 companies. FAANG applications open Aug-Sept; interviews run Sept-Dec.
Junior summer
/ top-tech summer internship. Conversion: 60 to 80% at FAANG. This is the main full-time pipeline.
Senior year
Full-time offer finalization. If you didn't get , target high-growth public companies (Stripe, Databricks, Snowflake) or strong AI-tier startups.
AI lab track (OpenAI, Anthropic, DeepMind)
Publish at NeurIPS / ICML / ICLR during undergrad or MS, or ship an obviously impressive open-source project / demo. Most hires have a , but not all.
Where this leads
Where this leads
Exit opportunities by stage
Software is less a 'pipeline' and more a career-long menu of moves — , startups, founding, or pivoting into / exec roles all stay open at every stage.
After 2 to 4 years (SWE II / III)
lateral
$280k to $450k total comp
Most common move. A senior role at Google or Meta often pays 40 to 80% more than the same title elsewhere.
Startup Engineering ( to C)
$180k to $260k + meaningful equity
Equity lottery ticket. Cash comp takes a 20 to 40% hit vs. big tech but upside is real.
Founding engineer
$140k to $200k + 0.5 to 2% equity
or seed startups. Lowest cash pay, highest equity. Most founding engineers get nothing; a few make $10M+.
After 5 to 8 years (Senior / Staff)
Staff / Principal at
$500k to $900k+ total comp
L6 / L7 levels. This is where most engineers top out and settle in.
AI lab researcher / engineer
$700k to $2M+ (varies wildly)
OpenAI, Anthropic, xAI, DeepMind. Highest-paying engineering seats in the world right now.
Engineering Manager / Director
$400k to $800k total comp
Management track. Pays similarly to senior IC but different skill set — people-first.
Founder / CTO at startup
$100k to $200k + 10 to 40% equity
Founding a startup. Most common after 5 to 10 years at .
After 10+ years (Staff / Principal / Director)
VP Engineering
$500k to $1.2M+ with equity
VPE roles at scale-ups or mid-stage startups. Equity can be life-changing on a good exit.
CTO at startup
$250k to $400k + 2 to 10% equity
Early-stage CTO roles. Lower cash, highest equity upside.
Investor / advisor
Highly variable
Angel investing, scout checks, or joining a fund as an Operating Partner.
If it doesn't work out
The honest back-up
If or AI labs don't come through
Top-tier tech is competitive, but the skills that got you close to a offer are in demand almost everywhere. The cash ceiling in software is broad - you don't need a logo to reach it.
- Mid-size / high-growth tech (Stripe, Databricks, Snowflake, Ramp, Figma, Anthropic Eng) - often higher total comp than the typical offer.
- Fintech and quant firms (Jane Street, Citadel, Two Sigma, HRT, Jump) - the highest cash comp in tech for engineers who enjoy low-latency or quantitative work.
- Enterprise software (Salesforce, Oracle, ServiceNow, Atlassian) - stable comp, solid equity, meaningfully better lifestyle than .
- Defense-tech (Anduril, SpaceX, Palantir, Shield AI) - pre-IPO equity that has recently rivaled outcomes, with meaningful mission.
- Founder track - joining or starting a -stage or seed-stage company; the equity math on a successful exit frequently dwarfs 10 years of .
Who hires for this
Example companies & employers
FAANG / top-tier tech (highest base + stock)
AI labs (highest pay in the industry)
High-growth public
Stable enterprise
Startups (YC / Seed / Series A to C)
The playbook
How to actually break in
Build real things people can actually use
Hiring is portfolio-driven. Ship 2 to 3 non-trivial projects on GitHub: a full-stack app with a database, a browser extension that solves something real, or contributions to an existing open-source project. Tutorials don't count. A working portfolio beats a CS degree from a no-name school.
Get the fundamentals rock solid
You need data structures, algorithms, and system design to pass interviews. Grind LeetCode for 200 to 400 problems over 3 to 6 months. Read 'Designing Data-Intensive Applications' once you have real job experience. This is the skill most self-taught folks never cross.
First job is the hardest
Aim for any real engineering role at a real company, even if it's a regional bank or a no-name SaaS. Title and pay matter much less at this stage than the word 'Engineer' on your resume with commits to a production codebase. Average time to first job for self-taught: 6 to 18 months of serious effort.
For : practice the loop, then interview everywhere at once
interviews are 4 to 6 rounds: 2 to 3 coding, 1 to 2 system design, 1 behavioral. Prep 200+ LeetCode mediums and 10+ system design case studies. Apply to 6 to 10 FAANG-adjacent companies simultaneously. Offers get used to negotiate against each other, often adding $50k to $150k to your first-year comp.
For AI labs: publish or build something people recognize
OpenAI, Anthropic, and DeepMind hire mostly from (a) published ML researchers with top-conference papers (NeurIPS, ICML, ICLR), or (b) engineers who shipped something obviously impressive (popular open-source ML library, viral demo, infrastructure used by others). A good ML plus published work is the most common path in. The 'resume-only' path works for ~0 people.
Switch companies every 2 to 3 years for the first 8
Internal promotions cap at ~10% raises. External moves can be 30 to 60%, and the stock refresh resets on every offer. Most people who hit $500k+ total comp by 30 did it through 3 to 4 company switches, not loyalty. Stay long enough to earn the title, then jump.
Specialize once you're senior
At the senior level (~5 years in), pick a lane: ML, infrastructure, security, distributed systems. Generalists plateau around $220k. Specialists, especially in ML and infra, routinely clear $500k+ at top companies, and AI-lab researchers can clear $1M+.
Realistic timing
How long each phase takes
6 to 18 months of self-study to first junior job. 3 to 5 years to senior engineer. 7 to 10 years to staff or principal. Elite path: junior → senior in 5 years → L6 Staff by year 10 → $500k to $900k total comp.
Avoid these
Common, expensive mistakes
- 01Spending $100k+ on a private CS degree when a state school or bootcamp plus a strong portfolio would perform identically.
- 02Doing tutorials for 2 years without shipping anything original. Projects teach 10x what courses do.
- 03Refusing to LeetCode because 'it's not real engineering.' You can't pass the interview without it.
- 04Going to a bootcamp and expecting immediately. FAANG almost never hires bootcamp grads directly; you usually need 2 to 4 years at a smaller company first.
- 05Staying at a no-growth company for 5+ years waiting to be recognized. Nobody is coming to rescue your career.
The honest version
Who actually makes it
People who can sit with frustration for hours debugging something nobody else understands. Enjoying the puzzle matters more than being a 'math person.' The drop-out rate in this field is high, almost all of it from people who realize they don't actually enjoy the work itself.
Routes in
Education paths & payback
Multiple real paths into this career, each with different odds, costs, and risks.
Smart start
Top public CS program (in-state)
Software is the rare high-paying career where a state school is genuinely the smart path. About 9 public universities have computer-science programs that big tech companies recruit at as hard as MIT and Stanford: Berkeley, Michigan, UIUC (Illinois), Georgia Tech, UT Austin, UCLA, Wisconsin, Purdue, and University of Washington.
Placement reality
Berkeley, UIUC, Georgia Tech, and UT Austin each send 100-300+ grads/yr to (Facebook/Meta, Apple, Amazon, Netflix, Google) and other top tech companies. AI labs (OpenAI, Anthropic, Google DeepMind) recruit from this list just as hard as from MIT or Stanford.
Total cost
$60k
Length
4 yrs
How we calculate it: total education cost ÷ (15% of first-year pay put toward paying it back).
Smart start
Top private CS school
MIT, Stanford, CMU (Carnegie Mellon), Caltech, Princeton. Highest density of top-tech and AI-lab hires per student. Costs a lot more than the state schools, but starting salaries are high enough to absorb it.
Placement reality
MIT and Stanford CS each send 30-50% of their graduating class to top tech companies or top startups. AI lab hiring is densest here.
Total cost
$280k
Length
4 yrs
How we calculate it: total education cost ÷ (15% of first-year pay put toward paying it back).
Standard route
Self-taught + portfolio
About 6-12 months of focused practice. You'll need a strong GitHub profile (an online portfolio of your code) and real shipped projects you can show. Works best for web/app development at startups; much harder at .
Placement reality
More common than people realize — about 25% of working software engineers don't have a CS degree. But and AI labs increasingly require degrees, so most self-taught engineers start at startups or smaller tech companies first.
Total cost
$2.0k
Length
2 yrs
How we calculate it: total education cost ÷ (15% of first-year pay put toward paying it back).
Standard route
Coding bootcamp + strong portfolio
A 'bootcamp' is a 3-6 month intensive coding program (App Academy, Hack Reactor, etc.). Realistic path for entry-level web / app development jobs at startups or mid-size tech companies.
Placement reality
Bootcamp grads rarely land at or AI labs directly. Realistic first job: $90-130k at a startup or mid-size tech company, then climb. AI/ML research roles essentially require a degree.
Total cost
$15k
Length
1 yrs
How we calculate it: total education cost ÷ (15% of first-year pay put toward paying it back).
Standard route
Any other CS degree
Any state school's computer-science program. The career still works, but landing at or AI labs is meaningfully harder than from the top public CS programs.
Placement reality
Most grads land at startups, mid-tier tech companies, or IT-style roles first. Jumping up to after 2-4 years of solid experience is the most reliable upgrade path.
Total cost
$50k
Length
4 yrs
How we calculate it: total education cost ÷ (15% of first-year pay put toward paying it back).
The real cost will be different for you
The numbers above are medians from the U.S. Bureau of Labor Statistics. The price you actually pay depends on financial aid, scholarships, family help, state residency, and which schools accept you. A "$250k" private university is often $60k to $120k after aid. A "cheap" community college can still be out of reach without a plan. Use these as a starting point, then get real quotes from the specific schools you're considering.
Location matters
How pay changes by metro
The same job in different cities can pay 2 to 3 times more. Location may matter more than everything else.
| Metro | Pay at Senior Engineer |
|---|---|
San Francisco Bay Area, CA FAANG + highest-paying startups; staff engineers clear $700k+. | $319k |
Seattle, WA Amazon, Microsoft, Meta offices. | $286k |
New York City, NY Fintech (Two Sigma, Jane Street) + big tech satellites. | $275k |
Los Angeles, CA | $242k |
Boston, MA | $242k |
Austin, TX Big tech satellites, no state tax. | $231k |
Denver, CO | $220k |
Remote (US national) Many mid-size tech companies pay a flat national rate. | $209k |
Chicago, IL | $202k |
Atlanta, GA | $194k |
Phoenix, AZ | $180k |
Rural / small metro Limited to small local shops or remote roles. | $154k |
How to read this: Pay is US median for Senior Engineer scaled by BLS metro wage data. Cost of living index is C2ER (US avg = 100). A means that metro isn't just paid better. It's essentially required to work at the top of this field.
Start now
What you can do in high school
Classes & electives
- AP Computer Science A
- Calculus
- Statistics
Things to try outside class
- Build and ship something real: a website, game, or app
- Do Advent of Code or LeetCode easy problems
- Contribute to open source, even tiny PRs
- Try CS50 (Harvard's free intro). It's famous for a reason
Read, listen, watch
The cheapest way to find out whether you actually like a field is to read what people in it read.
- 'The Pragmatic Programmer' by Hunt & Thomas
- 'Soul of a New Machine' by Tracy Kidder (Pulitzer Prize, still the best book on what engineering feels like)
- Paul Graham's essays at paulgraham.com (free, short, written for young programmers)
- Lex Fridman and The Changelog podcasts
None of this is required. It's just the lowest-friction way to find out if you actually like this kind of work before you commit money or time.
Salary and progression source: Industry compensation surveys and public offer data (BLS does not track these roles separately). Figures are medians or typical ranges in today's dollars; real outcomes vary by employer, city, and specialty.