How the questions work, and how they are marked.
What you are asked in each round, and how every answer is marked.
From an online paper to an offline final.
Round 1 is sat on this platform; Rounds 2 and 3 are run in person by the organisers.
- 1On this platform
Round 1 · Online
Open to 4,000+ candidates
One-hour modules across analog, digital, DSP and embedded, with fastest-finger-first sections. Sat in the desktop app on your own laptop.
- 2In person
Round 2 · Regionals
Top 1,500
Proctored, computer-based testing at partner campuses: a deeper technical evaluation.
- 3In person
Round 3 · Offline finals
Top 70–80
In Chennai: handwritten design contests and sponsor-run technical events.
Two contests, one for interns and one for placements.
Exact dates are on each contest’s own page once it is published.
Intern hiring contest
- Registrations open mid-March
- Contest in April
Placement hiring contest
- Registrations open mid-September
- Contest in October
How the Round 1 paper is put together.
Aptitude for everyone, then a question section and a build problem for each domain you choose.
- 01
Aptitude
EveryoneMultiple choice, sat by every candidate.
- 02
Choose your domains
One or more of fourAnalog, digital, signal processing, embedded: one, or several.
- 03
Domain questions
One set per domainMultiple choice in each domain you chose, marked against the key.
- 04
Built answers
One per domain you choseA circuit to design or code to write, one per domain. Built answers are run, not read.
Pick the domains you would take
- 01AptitudeMultiple choice
- 02DigitalMultiple choice
- 03EmbeddedMultiple choice
- 04Digital buildDesign or code
- 05Embedded buildDesign or code
5 sections: aptitude, then questions and a build problem in Digital, Embedded.
An illustration. The real choice is made in the app, during the paper.
The rules of the paper
- Parts and sections
- Sections are grouped into parts. Each part has its own time box, instructions and scoring rule, stated with the paper.
- Choosing your domains
- After the aptitude section you choose one or more of analog, digital, signal processing and embedded. Each brings a question section and a build problem. The choice is made in the app, during the paper.
- Inside a part
- Move freely between its sections; every answer is saved the moment it changes.
- When a part ends
- It closes for good. Leftover time does not carry into the next part.
- Fastest finger first
- Short questions of the ordinary types, marked the ordinary way. No clock per question; speed only breaks ties.
- Built sections
- A digital build is set as Verilog to write or a circuit to draw. The other domains set a design or coding problem of their own.
- Where you sit it
- In the desktop app on your own laptop. The clock is the platform’s. If your connection drops, reopen the app and continue with the time left.
What you see, what you do, how it is marked.
Question sections are marked against a key; build problems by running what you built.
Question sections
Aptitude, and one for every domain you takeMultiple choice. A paper may also use the three other quick types below.
Multiple choice
- What you see
- A statement, an image when there is one, and options. A section is single-correct or multi-correct, never mixed.
- How it is marked
- Against the key. Marks and negative marks per question; a multi-correct section may give partial credit.
True or false
- What you see
- A statement and two fixed options.
- How it is marked
- Against the key.
Fill in the blank
- What you see
- A statement with a gap, and a text box.
- How it is marked
- Against the accepted answers the setter provides.
Numerical
- What you see
- A statement and a box for a value, sometimes with a unit.
- How it is marked
- Against an expected value, within a tolerance the setter sets.
Build problems
One for every domain you takeA circuit to design or code to write. A digital build is set as Verilog to write or a circuit to draw; the question says which.
Verilog
- What you see
- The statement and a code box holding the module with its ports declared. You write the inside and press Run.
- How it is marked
- Compiled and driven by a hidden testbench. A compile error shows the compiler’s message; a failure shows the waveform at the first mismatch, so you can fix it and run again.
Drawn circuit
- What you see
- A drawing area with the question’s inputs and outputs fixed at its sides. You drag in gates, multiplexers and flip-flops and wire them.
- How it is marked
- By the same hidden testbench, on behaviour alone: any circuit with the right outputs is correct. A drawing that cannot be simulated is refused with the reason.
Analog, signal processing and embedded
- What you see
- A design or coding problem set for that domain.
- How it is marked
- As the contest’s paper states for that domain.
How a score is built, and how a rank is decided.
Nothing in Round 1 is marked by hand.
- Per question
- Marks, and optionally negative marks, set per question and shown with the paper.
- Built answers
- Run against a hidden testbench; a pass earns the marks. The number of runs, the time limit per run and any penalty for a wrong attempt are set by the contest and told to you before you start.
- A guarantee before you see the paper
- Every built question has a reference solution, and a contest cannot be published until each one passes its own testbench.
- Per part
- The sum of its sections, or the best K of them, as the paper states.
- The domains you take
- How different domains count towards one total is part of the scoring rule stated with the paper.
- Total and rank
- Your total is the sum of your parts. Rank is by total; ties are broken by total time.
- Adjustments
- A wrong question is given full marks to everyone or dropped from every total. Penalties and disqualification are applied by an organiser with a recorded reason.
- The leaderboard
- Not public until the organisers publish it. It then shows rank, username, score and penalties; whether your name and email appear depends on a per-contest setting stated before you register. See what is public.
- Going forward
- The top of the Round 1 board, up to 1,500, goes to Round 2; the top 70–80 from there to the finals.
See both kinds of marking.
A digital question marked against the key, and a build you can run and watch fail.