Study Buddy turns "prepare for top tech interviews" into a plan you can follow around a full-time job. Tell it the level you are aiming for, the companies, how many hours you really have and how strong you are in each area, and it builds your week, session by session, then adjusts to how each session actually went.
What it does
Five tracks, weighted by level and target company: DSA, LLD and machine coding, HLD and system design, behavioural, and AI-assisted coding.
System design follows Alex Xu's System Design Interview, volumes 1 and 2, chapter by chapter: design it yourself first, then read, then write down what you missed.
Spaced review: a topic you struggled with comes back the next day, one you solved comes back after a growing gap.
An error log that records why each attempt failed, so practice goes where you actually lose points.
A timeline check that says whether your hours fit your target date, and what would make them fit.
Text mock interviews for behavioural, system design and LLD rounds, graded against a written rubric, plus grading for AI-assisted coding sessions.
A resume review, an applications and referrals tracker, and offer comparison across currencies with negotiation prep.
An ATS-style resume check with no AI: upload a PDF or DOCX and see whether it parses as real text with standard headings, contact details and dates, and which skills from a job description it mentions, from a fixed list of common engineering terms.
Prep notes for 20 companies, and a focus mode for the 21 days before an interview date in your tracker: more review time and more of that company's round formats.
How it works
1Sign in and set your level, target companies, weekly hours and study days.
2Rate yourself in each track; the plan leans towards your weakest.
3Open Today, do the sessions, and log how each one went.
Built on how learning works, not on personality types
The plan does not sort you into a learning style - research has not found that teaching to a style helps. It uses what does: spaced repetition, re-solving from a blank page instead of rereading, mixing topics once the basics are in, and spending most of your time on your weakest areas.
What top tech interviews test in 2026
A typical loop has two or more coding rounds, a design round and a behavioural round. Indian product companies add a 90-minute machine coding round graded on OOP, SOLID and design patterns, and companies such as Meta, Google and Canva now run rounds where you code with an AI assistant and are graded on how well you direct and verify it.
How the plan decides what you study
Each level starts from its own split of study time. SDE-1: 55% DSA, 15% LLD, 5% system design, 15% behavioural, 10% AI-assisted coding. SDE-3 and above: 35% system design, 20% DSA. Target companies known for machine coding move time to LLD, companies with AI-assisted rounds raise that track, and your self-rating bends the split again: a track rated 1 out of 5 gets 1.4 times its share, one rated 5 gets 0.8 times. Reviews follow a ladder of 1, 3, 7, 14, 30 and 60 days: solving on your own moves a topic one rung up, needing a hint keeps the gap, and a fail sends it back to 1 day. Reviews take at most 40% of a day (50% in the 21 days before an interview), so new material keeps moving.
Where AI is used, and where it is not
The schedule, the timeline check, the ATS check and the offer maths are plain arithmetic with no AI, so the same inputs always give the same answer. Google's Gemini 2.5 Flash runs the coach chat, the mock interviewer, grading of mocks and of your session notes, the pattern check, the resume review and negotiation prep. Mocks are scored 1 to 4 per criterion on a fixed rubric, and the model's reply is checked before it is saved. Email addresses and phone numbers are masked before a resume reaches the model. Limits per account: 30 coach messages and 60 mock interview turns an hour, 20 mock gradings and 10 resume reviews a day.
How the offer comparison adds up
Every offer is spread over four years. A year's total is base salary + expected bonus + other yearly value + that year's share of the equity grant; the joining bonus and relocation count in year 1 only. Vesting is up to four yearly percentages adding up to 100, such as 25,25,25,25 or 5,15,40,40. Amounts in other currencies are converted with exchange rates you can edit (the starting values are rough, so enter today's rates), and the 4-year average is divided by the offer's cost-of-living factor: 1 for your current city, 1.5 where living costs 50% more. The adjusted figure is also shown out of 100 against the best offer. Equity counts at grant value with no rise or fall assumed, no tax is deducted, and no AI is involved.
What the ATS check looks at
There is no universal ATS score: systems such as Workday, Greenhouse and Lever parse a resume into fields that recruiters then search by keyword. So the check reports only what it can read off your resume, uploaded as a PDF or DOCX of up to 5 MB or pasted as text, and uses no AI. It fails an upload that gives under 300 characters of text (usually a scan) and a resume with no email address. It warns when no English heading for Experience, Education or Skills sits on its own line, when it finds no phone number or no dates for your roles, when dates mix formats such as "Jan 2022" and "01/2022", when a PDF runs past two pages or the text is under 250 or over 1,100 words, when a DOCX uses tables or puts text in the page header or footer, and when icon-font symbols or garbled characters turn up.
Questions
Do I need an account?
Yes. Study Buddy keeps your plan, progress and error log, so it is for signed-in users only. You can delete all of it at any time from the setup page.
Where do I solve the problems?
On LeetCode or wherever you already practise. Study Buddy links each problem set and you log how the attempt went.
Which book does the system design track use?
Alex Xu's System Design Interview, volumes 1 and 2. Each chapter is one session, plus extra designs from recent interview reports.
What if I miss a day?
Today's plan is rebuilt from your results, so unstarted topics come up again on your next study day. Reviews that fell due while you were away wait for you; once more than two days' worth have piled up, they may take up to 60% of a day, oldest first, until the backlog is cleared. You can also mark days off or light days ahead of time, and the timeline check counts them.
Which companies does it have notes for?
Google, Meta, Amazon, Microsoft, Apple, Netflix, Uber, Atlassian, LinkedIn, Canva, Shopify, Rippling, Salesforce, Adobe, Flipkart, Razorpay, Swiggy, Zomato, PhonePe and CRED. The notes cover widely reported round formats only, and formats change, so confirm yours with the recruiter.
Why do two offers with different vesting show the same average?
Because the same grant adds the same total over four years. Take a base of 100,000, an equity grant of 200,000 and a joining bonus of 20,000. Vesting 25,25,25,25 gives 170,000 in year 1, then 150,000 in years 2 to 4; vesting 5,15,40,40 gives 130,000, 130,000, 180,000 and 180,000. Both average 155,000, so compare the Year 1 and Year 2 columns, especially if you might leave before the grant fully vests.
Does the decision score use the salary figures?
No. The decision matrix uses only the 1-to-5 score you give each offer on eight factors (compensation, learning and growth, team and manager, the work itself, work-life balance, location or remote, stability, brand and next move), each multiplied by your 1-to-5 weight for that factor, and shows the result out of 100. Score compensation yourself after reading the 4-year table.
How is the keyword match percentage worked out?
Paste a job description and the check finds which of 87 common engineering skills it names, counting aliases such as k8s for Kubernetes and Postgres for PostgreSQL, then shows what share of them your resume also names, with the matched and missing skills listed. A description naming 8 of those skills, 6 of which are in your resume, scores 75%. Below three recognised skills there is no score, and terms outside the list, such as in-house tools, are not checked.