Article details
Every placement season we hear the same line from final-year ECE students: "core jobs don't exist anymore, everybody just shifts to IT."
It is half true. Campus drives are still dominated by service companies hiring for roles that have nothing to do with electronics. But the other half of the picture is very different in 2026. Fabs are coming up in Gujarat and Assam, design services firms are hiring in Bengaluru, Hyderabad, Noida and Pune, and EV and defence electronics are pulling in hardware people faster than colleges can produce them. The openings are there. What is missing, usually, is a candidate who can prove they can do the work.
That last part is the whole article. Here is how to get on the right side of it.
Pick one track, then go deep
ECE is not one field. It is five or six fields wearing the same degree certificate, and the biggest mistake students make is trying to stay "open to everything" until the last semester. Recruiters can smell that instantly. A student who has done a bit of Arduino, a bit of MATLAB and a bit of Java looks weaker than one who has done nothing but embedded firmware for eight months.
The main tracks worth committing to:
Digital design and VLSI. RTL design, functional verification, physical design, static timing analysis, design for testability. The largest volume of core hiring in India right now, mostly through design services companies and captive centres of the big semiconductor firms.
Embedded systems and firmware. Microcontrollers, drivers, RTOS, board bring-up. Demand comes from automotive, medical devices, industrial automation, consumer products and defence.
RF, antennas and wireless. Smaller intake, higher barrier. 5G infrastructure, satellite communication, radar, and the antenna and front-end work inside every connected device.
Signal processing and edge AI. Filtering, estimation, audio and vision pipelines, and increasingly running small neural networks on microcontrollers instead of the cloud.
Power electronics and instrumentation. Converters, motor drives, battery management, SiC and GaN devices. EV manufacturers and their suppliers are the obvious employers, but so are solar, rail and industrial drives companies.
You are allowed to change your mind. What you are not allowed to do is stay undecided for three years and then hope a company decides for you.
The fundamentals interviewers keep returning to
Whatever track you choose, a technical interview will circle back to a fairly small set of topics. I have sat in on enough of these to say the pattern is remarkably consistent.
Digital electronics comes up first, almost always. Flip-flops versus latches. Setup and hold time, and what actually happens when you violate hold. Metastability, and why a two-flop synchroniser reduces it rather than eliminating it. Clock domain crossing. FSM design, and whether you can draw a Mealy and a Moore version of the same machine. Blocking and non-blocking assignments in Verilog, which is possibly the single most-asked question in Indian VLSI interviews.
Analog fundamentals come next. Op-amp configurations, biasing, MOSFET regions of operation, small signal models, and the basic behaviour of an RC circuit under a step input. You do not need to be an analog designer to be asked these.
Then C programming, if you are anywhere near embedded. Pointers, pointer arithmetic, volatile and const volatile, memory sections, stack versus heap, bit manipulation, structure padding, function pointers, and what an ISR should never do. Interviewers love asking you to set, clear and toggle a specific bit without disturbing the others. Practice writing it on paper.
Signals and systems, communication theory and networks fill the rest. Sampling and aliasing, convolution, FFT versus DFT, modulation schemes and where each one is used, and for wireless roles, transmission line behaviour and the Smith chart.
And yes, you will also face coding rounds. Many core companies screen with the same platforms IT companies use. Arrays, strings, linked lists, recursion, sorting, basic complexity analysis. You do not need to be a competitive programmer, but you cannot walk in blank either. Two months of steady practice is usually enough for the level these rounds are set at.
Build two projects that can survive cross-examination
Not eight. Two, maybe three, that you can defend for forty minutes.
The difference between a project that helps you and one that wastes your time is whether it had constraints. A line-following robot built from a YouTube tutorial teaches nothing that shows up in an interview. The same robot, where you had to hit a specific loop time, debug an I2C bus that kept locking up, and explain why you moved from delay-based timing to a hardware timer, is a genuine talking point.
Some project directions that hold up well:
VLSI: a pipelined RISC-V core in Verilog, or a UART or SPI controller written from scratch and verified with a SystemVerilog testbench. Push it through synthesis. If you can take a small design through OpenLane on the SkyWater 130nm PDK and produce a GDSII, you are ahead of most applicants.
Embedded: a sensor node on an STM32 or ESP32 with FreeRTOS, multiple tasks, a real communication protocol and proper power management. Bare-metal register-level code for at least one peripheral, because it proves you can read a datasheet.
RF and wireless: a microstrip patch antenna designed in HFSS or CST, fabricated if your lab allows, with measured versus simulated return loss. The gap between the two is the interesting conversation.
Signal processing and edge AI: keyword spotting or vibration-based fault detection running on a microcontroller with TensorFlow Lite Micro or CMSIS-DSP. Report the memory footprint and the inference time.
Power electronics: a buck or boost converter designed, simulated in LTspice or PLECS, then actually built, with efficiency measured across the load range and a plot to show for it.
Put everything on GitHub. A README that states the specification, the design decisions, what failed, and how you fixed it is worth more than the code itself. Hiring managers read those.
Learn a few tools properly
Listing eleven tools on a resume signals that you have touched all of them and mastered none. Depth in two or three is far more persuasive.
Digital and VLSI people should be fluent in Verilog and SystemVerilog, comfortable in a simulator, and able to script in TCL and Python. Familiarity with Synopsys, Cadence or Siemens EDA flows helps; if your college does not have licences, the open-source flow is a legitimate substitute and shows initiative.
Embedded people need Git, a real IDE or toolchain, a debugger they can use without hand-holding, and hardware instruments. Being able to read a signal on an oscilloscope or a logic analyser and say what is wrong with it separates you from candidates who only know simulation.
Everyone should have Python. Not for the sake of it, but because test automation, data logging and result analysis are part of almost every hardware job now.
Your resume gets about seven seconds
One page. Projects above internships if you are a fresher with limited work experience, and both above the "hobbies" section that nobody reads.
Write bullets that carry a number and a technology. "Reduced boot time from 1.8 s to 640 ms by moving initialisation to a deferred task" tells me something. "Worked on embedded systems" tells me nothing. Drop the adjectives about being hardworking and passionate. Keep the GitHub link, keep the CGPA if it is decent, and match the vocabulary of the job description closely enough for the screening software to find you.
One resume per track. If you are applying to both VLSI and embedded roles, you need two documents. This is not dishonesty, it is emphasis.
What the process actually looks like
Most core companies run something close to this: an online aptitude and technical test, a first technical interview on fundamentals, a second and deeper interview built around your projects and one or two open-ended design questions, then HR.
The second round is where offers are won and lost. Expect to be asked to design something on the spot. Design a traffic light controller as an FSM. How would you debounce a switch, in hardware and in software. Design a 4-bit synchronous counter and then tell me its maximum operating frequency. How would you talk to twelve sensors with an MCU that has two SPI peripherals. If your project used an I2C sensor, why I2C and not SPI, and what happens on the bus when a slave stops responding.
When you do not know something, say so and then reason out loud. Interviewers are testing your approach at least as much as your recall. Silence or a confident wrong answer both hurt more than an honest "I have not worked with that, but here is how I would start."
Where the openings actually are
Campus placement is one channel, not the only one. Some of the better routes:
Design services and product engineering firms hire in large numbers and give freshers real exposure. Many people enter the semiconductor industry through them rather than directly into a product company.
Internships convert. A six-month internship in your final year, even an unpaid one at a small hardware startup, is often a better path to a job offer than a dozen applications.
PSUs through GATE remain a strong option if you want stability, and the preparation overlaps almost entirely with what interviews test anyway.
Referrals move faster than portals. Being active and specific on LinkedIn — posting what you built, commenting on other engineers' work — genuinely leads to conversations. Recruiters look at profiles that show output.
And keep an eye on the smaller hardware startups. Fewer applicants, broader work, and you learn far more in the first year.
A six-month plan if you are starting from zero
Month one and two: rebuild fundamentals in your chosen track and start structured coding practice. Month three and four: first serious project, documented as you go, plus tool fluency. Month five: second project, resume, GitHub cleanup, LinkedIn. Month six: mock interviews, aptitude practice, and applications going out every week rather than in one anxious burst.
Six focused months change your position completely. Three years of unfocused effort will not.
The mistakes I see repeatedly
Chasing certificates instead of finishing projects. Applying only during campus season. Preparing for coding rounds and neglecting core, or the reverse. Copying a project without understanding it, then being unable to explain why a particular capacitor value was chosen. Waiting until eighth semester to begin. And treating the degree as the qualification, when in this field it is only the entry ticket.
One last thing
Core electronics hiring rewards evidence. Nobody in a technical interview cares what you have read. They care what you have built, broken, debugged and understood.
Pick your track this month. Start the first project this week. The rest tends to follow.
