13 companies graded on whether they want new grads

I'm going to tell you something that nobody told me when I was graduating: the company you apply to matters more than how hard you grind LeetCode.

Not because some companies are "easier." Because some companies have interview loops that are structurally designed for people with industry experience — and if you're a new grad with zero years of experience, you're not being evaluated on a level playing field. You're being asked to demonstrate signals you haven't had a chance to build yet.

We track 13 of the most sought-after tech companies in the StrongYes dossier store. For each one, we've read the candidate reports, mapped the interview loops, pulled the comp data, and written a dedicated "new grad entry point" section. This post is the synthesis: a tier list that tells you where your application has the best structural chance of converting to an offer.

The methodology is simple. We scored each company on four dimensions:

1. Hiring volume — Does the company hire new grads at scale, or are junior spots rare?
2. Interview accessibility — Does the loop skip system design at entry level? Are the rounds well-documented and predictable?
3. Cooldown leniency — If you bomb it, how long until you can try again?
4. New-grad accommodations — Does the company calibrate expectations for zero-YOE candidates, or do they apply the same bar regardless?

Here is how the 13 companies sort out:

Diagram

Copy source

Rendering diagram...

Now the details. Every company gets one "killer detail" — the specific structural feature that will decide whether your application converts.

---

Tier 1: GO FOR IT

These companies actively hire new grads at volume. The interview loops are well-documented, skip system design at entry level, and give you a realistic shot even with zero industry experience.

Microsoft — L59 (SDE) — ~$160K TC

The friendliest door in Big Tech.

Microsoft has the highest Glassdoor interview satisfaction in the entire dossier store at 64% positive. That's not a typo — nearly two-thirds of candidates say the experience was good. For context, the average across our 13 companies is around 40%.

The loop: phone screen with one coding problem, then a 4-5 round virtual onsite. Coding + behavioral. No system design at L59. The rounds are well-documented on Glassdoor and LeetCode Discuss, which means your prep is predictable.

The killer detail: no formal cooldown after rejection. Most Big Tech companies make you wait 6-12 months (Google is 12). Microsoft doesn't publish a mandatory cooldown. If you bomb it, you can potentially reapply much sooner than you'd expect.

The tradeoff: the lowest TC in Tier 1 at ~$160K. You're trading comp for accessibility. For a new grad who needs a foot in the door, that's often the right trade.

Amazon — L4 (SDE I) — ~$185K TC

The volume play.

Amazon is one of the highest-volume new-grad hirers in the store. They run massive hiring pipelines, and they let you interview with multiple teams simultaneously. If Team A passes and you don't love the role, you can match with Team B without re-interviewing.

The loop: online assessment → phone screen → 4-round virtual onsite (2 coding + 1 system design lite + 1 behavioral). The behavioral round is ALL about Leadership Principles — Amazon's 16 principles are the backbone of every interview decision.

The killer detail: Y1/Y2 RSU cliff. Amazon's $185K TC includes an unusually low Year 1 base + signing bonus, with RSUs backloaded to Years 3-4. Your first-year cash is significantly lower than the "TC" number suggests. Know this before you negotiate.

The tradeoff: Amazon's interview satisfaction is 48% positive — still below Microsoft. The LP-heavy behavioral round trips up new grads who haven't memorized and practiced all 16 principles with STAR-format stories.

Meta — E3 (Software Engineer) — ~$173K TC

The one that banned DP.

Meta made a structural decision that disproportionately helps new grads: they banned dynamic programming from E3 (entry-level) interviews. If you've been grinding DP problems because "that's what FAANG asks," Meta is telling you not to bother — at least not for them.

The loop: recruiter screen → 2 coding screens → onsite (2 coding + 1 behavioral + 1 system design). Wait — system design at E3? Yes, but Meta calibrates it significantly for entry level. You're not designing Netflix at scale. You're designing a simple feature with basic tradeoffs.

The killer detail: Meta is rolling out an AI-assisted coding round for SWE roles in 2026. Since October 2025, E6 and below engineers have been piloting AI tooling in onsite rounds. Early candidate reports suggest it will sit alongside the traditional rounds, not replace them (unlike DoorDash, which fully replaced its coding round with an AI session).

The tradeoff: Meta's interview loop gives hints. If you're stuck, the interviewer will nudge you toward the right approach. This is a genuine cultural choice — Meta wants to see how you respond to guidance, not whether you can solo a problem in silence. New grads who are good at collaborative problem-solving have an edge here.

Google — L3 (Software Engineer II) — ~$210K TC

The highest-volume hirer with the weirdest constraint.

Google hires more new grads than any other company in the store. The pipeline is enormous, the hiring committees are well-oiled, and the process is thoroughly documented.

The loop: phone screen (1-2 coding problems in Google Docs) → 4-5 round onsite (coding + behavioral). No system design at L3. Standard Big-O analysis expected.

The killer detail: you write code in Google Docs. Not an IDE. Not a shared code editor. Google Docs. No syntax highlighting, no autocomplete, no running your code. If you've only ever practiced in LeetCode's editor or VS Code, this will feel like coding with boxing gloves.

The tradeoff: Google's cooldown is 12 months — the longest standard cooldown in the store. If you interview and fail, you're locked out for a full year. This means your Google application is a one-shot per year. Time it carefully.

---

Tier 2: WORTH A SHOT

These companies hire new grads, but the process adds friction: team variance, unusual round formats, or a higher bar than Tier 1. Navigable with targeted prep.

Apple — ICT2 (Junior Software Engineer) — ~$178K TC

The wildcard.

There is no single "Apple interview." Two new grads applying to different teams can get completely different loops — different number of rounds, different question types, different emphasis on coding vs. design vs. domain knowledge.

The loop: recruiter call → technical phone screen → 4-6 round onsite. The exact composition depends entirely on the hiring team. Some teams skip system design for ICT2. Some don't. Some add a domain-specific round (e.g., GPU knowledge for the ML infra team).

The killer detail: ask the recruiter what the loop looks like. At every other company, you can predict the loop from Glassdoor. At Apple, Glassdoor is unreliable because loops vary by team. The recruiter is your best source. Ask explicitly: "How many rounds? Is there system design? What should I focus on?"

The tradeoff: Apple's secrecy culture extends to interviews. There's less public prep material than for any other Tier 1/2 company. You'll rely more on your recruiter relationship and less on LeetCode Discuss.

DoorDash — E3 (Software Engineer) — ~$173K TC

The AI-native interviewer.

DoorDash is the only company in the store that has fully replaced its traditional coding round with a 60-minute AI-assisted engineering session. Cursor, Claude Code, Copilot — all allowed, all expected. "Prompt engineering ability" is an explicit graded dimension.

The loop: HackerRank OA → phone screen → AI-assisted working session → system design → behavioral. The AI session starts with a realistic DoorDash-flavored starter project (order dispatch, menu composer, etc.).

The killer detail: DoorDash explicitly grades you on how well you use AI tools. Candidates who avoid AI because they think it's "cheating" are penalized. Candidates who use AI to ship more features, catch more bugs, and iterate faster are rewarded.

The tradeoff: this is a genuinely new interview format. There's less historical prep data than for any traditional coding round. You need to practice with AI coding tools before the interview — not just install them the night before.

Stripe — L1 (Software Engineer) — ~$217K TC

The debugging interview.

Stripe doesn't do LeetCode. Their most distinctive round is a live bug bash in a cold codebase: you open a project you've never seen, read through unfamiliar code, find the bugs, and fix them. This tests a fundamentally different skill than algorithm interviews.

The loop: recruiter call → HackerRank OA → 2-part phone screen → 4-round onsite (coding + debugging + system design + integration). The debugging round is where most new grads either shine or struggle — it rewards candidates who are strong readers of code, not just writers.

The killer detail: Stripe pays $217K TC at L1. That's the third-highest new-grad comp in the store (behind Databricks at $252K and OpenAI at $249K), and you get it at a company with a friendlier interview process than either of those.

The tradeoff: the debugging round has no equivalent at other companies. You can't prep for it by grinding LeetCode. You need to practice reading and debugging unfamiliar codebases — open-source projects, code review exercises, or Stripe's own prep resources.

Uber — L3 (Software Engineer I) — ~$165-204K first-year TC

The one that makes new grads do design.

Uber is the only Tier 2 company that includes a low-level design / OO round for new grads. Most FAANG peers skip design entirely at entry level. Uber doesn't.

The loop: HackerRank OA → phone screen → 4-round virtual onsite (2 coding + 1 LLD/OO design + 1 behavioral). The LLD round asks you to design a class hierarchy or API for a realistic system (parking lot, ride-matching, etc.).

The killer detail: Uber's comp range is wide — $165K to $204K for the same L3 level, depending on team and negotiation. That's a $39K spread. Negotiation matters here more than at companies with tighter bands.

The tradeoff: the LLD round. If you've only prepped for algorithms, you'll be caught off guard. Practice designing clean class hierarchies, API contracts, and state machines.

---

Tier 3: UPHILL BATTLE

These companies either hire very few new grads, run interview loops designed for experienced engineers, or both. Not impossible — but go in with your eyes open.

Airbnb — G7 (Software Engineer) — ~$180K TC

Culture fit is the silent killer.

Airbnb has a culture fit gate that kills roughly 50% of candidates — regardless of technical performance. This isn't a soft signal. It's a hard gate: fail culture fit, and your technical scores don't matter. You will not receive an offer.

The loop: HackerRank OA (~20-25% advance rate — one of the tightest filters in the store) → recruiter → phone screen → 4-round onsite (coding + code review + system design + behavioral). Airbnb also has a unique code review round where you review a PR rather than write code.

The killer detail: runnable code or bust. Pseudocode is not accepted at any stage. Your code must compile and run. This is a higher bar than Google (which accepts pseudocode in Google Docs) or Meta (which allows partial solutions).

The tradeoff: Airbnb's centralized hiring model means you interview once and pick your team after. That's a genuine advantage — you're not locked into a team you don't want. But the 50% culture-fit kill rate means your interview odds are significantly lower than pure technical ability would predict.

Databricks — L3 (Software Engineer) — ~$252K TC

The highest-paying gauntlet.

Databricks pays the highest new-grad comp in the entire store at $252K TC. But only about 6% of applicants reach the onsite. That's not a typo — 94 out of 100 applicants are filtered before the final rounds.

The loop: recruiter → OA (~20-25% advance) → phone screen → 5-round onsite (2 coding + 1 system design + 1 concurrency/low-level + 1 behavioral). The concurrency round is unique to Databricks and tests multithreading, race conditions, and lock-free data structures — topics most new grads haven't studied deeply.

The killer detail: $252K TC as a new grad. If you can make it through the gauntlet, the payout is the highest in the store. That comp advantage compounds over time through refreshers and promotions.

The tradeoff: the concurrency round. Most CS curricula cover concurrency in one course. Databricks expects you to reason about thread safety, deadlocks, and memory models in a live interview setting. This is a senior-shaped question being asked of junior candidates.

OpenAI — L2 (Member of Technical Staff) — ~$249K TC

The most "studyable" — and the most chaotic.

OpenAI's interview has a known bank of approximately 8 recurring problems. Candidates report seeing the same KV Store, Excel, GPU Credit questions across multiple loops. This makes OpenAI the most "studyable" interview in the store — if you can get the questions, you can prep specifically for them.

The loop: recruiter → coding screen (four-gate progressive: 1 problem deepens through 4 levels, pass 2 = pass) → 4-5 round onsite (coding + system design on Excalidraw + project presentation + behavioral).

The killer detail: the process is widely described as "chaotic." Format changes mid-loop, communication gaps are common, and 95% of current employees joined after the November 2023 board crisis. The interview infrastructure is still catching up to the company's growth.

The tradeoff: AI is strictly prohibited despite OpenAI building GPT-4 and Codex. Practice raw coding — no Copilot, no autocomplete, no AI assistance. The irony is not lost on candidates.

Anthropic — MTS (Member of Technical Staff) — comp undisclosed (H-1B bases: $300K-$405K)

Values screening is the primary failure point.

At most companies, you fail the interview because of a coding question. At Anthropic, the most common failure mode is the values screening — and it happens before the technical rounds. Behavioral red flags that directly disqualify: lone wolf tendencies, arrogance, financial-only motivation.

The loop: two recruiter screens (values docs distributed on 2nd call) → coding screen (CodeSignal, 6-question progressive bank) → HM code review (analyzing existing codebases, not writing code) → 4-round onsite.

The killer detail: everyone gets the same title. A new grad and a co-founder can both be "Member of Technical Staff." The flat titling means your seniority is invisible externally — which can be a career-strategy consideration if you plan to job-hop within 2-3 years.

The tradeoff: 12-month cooldown, pre-IPO equity with no liquidity timeline, and a values bar that is genuinely different from "tell me about a time you showed leadership." Anthropic wants to know if you've thought about AI safety — not as a talking point, but as a personal conviction. Rehearsed answers are detected and penalized.

Netflix — E3 (Software Engineer) — ~$218K TC

The company that might not want you.

I'm going to be direct: Netflix historically does not hire many junior engineers. The company operated for 25 years without engineering levels. When they finally introduced levels in 2022, it was so controversial that it triggered significant attrition — only 24% of E5s planned to stay after the change. The culture is built around senior-density, and the interview reflects that.

The loop: recruiter → HM conversation → phone screen → 4-8 round virtual onsite. The round count alone tells you something — most companies top out at 5. Netflix can go to 8.

The killer detail: system design carries MORE weight than coding. This is the inverse of every other company in the store. For a new grad with limited design experience, this is a structural disadvantage that no amount of LeetCode grinding can overcome.

The tradeoff: if you get in, the comp is ~100% cash at $218K — no RSUs, no bonuses. Netflix's comp philosophy is "we pay you what you're worth in cash, and you decide what to do with it." But the keeper test means your job is never truly secure — Netflix evaluates whether they'd fight to keep you, and the answer determines your continued employment.

---

The Comp Cheat Sheet

Because you're going to ask:

---

So Where Should You Apply?

If you're a new grad in 2026, here's the honest answer:

Apply to all of Tier 1. Microsoft, Amazon, Meta, and Google are high-volume hirers with well-documented loops. Your odds are structurally the best here. Don't overthink which one to prioritize — apply to all four and see which pipeline moves fastest.

Pick 1-2 from Tier 2 based on your strengths. If you're a strong debugger, apply to Stripe. If you're comfortable with AI tools, apply to DoorDash. If you're good at OO design, apply to Uber. If you thrive in ambiguity, apply to Apple. Match your strengths to the round format.

Apply to Tier 3 only if you have a specific edge. Databricks if you've studied concurrency deeply. OpenAI if you've prepped the known question bank. Anthropic if you genuinely care about AI safety (not performatively — they'll know). Airbnb if your culture-fit stories are airtight. Netflix if you have 4+ years of experience and are reading this post ironically.

The single most important thing you can do today: stop grinding LeetCode hard problems and start researching the specific loop of the company you're applying to. The interview is not generic. Every company on this list has a unique format, a unique failure mode, and a unique thing that will surprise you. The dossiers exist for this reason.

Good luck. You're going to need preparation more than luck — but you already knew that.

---

Sources and Methodology

The tier rankings are editorial judgment based on the four dimensions above. The comp figures, interview-loop descriptions, and killer details are all traceable to specific per-company dossiers in the StrongYes store. Data drifts — we re-verify companies on a rolling basis, and the ranking order is the durable signal even when percentages shift.

• Levels.fyi — compensation medians by level and company. levels.fyi. The source of truth for the comp numbers in the cheat sheet. Filter to the specific level (Google L3, Meta E3, Stripe L1, etc.) and compare across companies — the relative ordering is stable even as absolute numbers shift.
• Glassdoor — company interview pages. glassdoor.com. Where the "interview satisfaction" figures come from (Microsoft 64%, Amazon 48%, etc.). Per-company URLs follow glassdoor.com/Interview/<Company>-Interview-Questions-E<numeric-id>.htm. Glassdoor is also the best source for recent candidate reports on round structure.
• StrongYes content directorate — per-company dossiers. Our internal dossier store tracks new-grad entry points, killer details, and tradeoffs across all 13 companies. The tier assignments are our editorial synthesis of those dossiers.

If a number in this post looks wrong, it's usually because Levels.fyi or Glassdoor has drifted since we last re-verified. Re-pull the company's page and assume the more recent number is correct.