Sougata Bhunia

India is at an inflection point. With the India Semiconductor Mission channelling public money into fabrication and packaging capacity, and companies such as Tata Electronics, Micron and CG Power building it out, demand for engineers who understand both the underlying physics and the process is growing faster than the country is producing them.

This matters to you specifically because the shortage is not at the top. There are plenty of brilliant graduates from the IITs. What the field is short of is people from ordinary colleges who took the physics seriously, learned to run instruments, and can be trusted with a process. That is a gap you can close by decision rather than by admission result.

The knowledge is also portable in a way the institution name is not. Diffraction, band structure, deposition, lithography and characterisation are the same in Taipei, Dresden and Sanand. Nobody at a bench has ever asked me where I did my B.Sc. once I could explain what I measured and why.

One caution: choose this field because the problems interest you, not because the market is warm. The training is long and the day-to-day is patient, careful work. If you do not enjoy chasing a discrepancy in a spectrum for two weeks, the salary will not carry you.

Specificity is everything. Open by naming one specific paper of theirs, in one sentence, and say what in it connected with your own work. That single sentence is what separates your email from the fifty templated ones they received the same week.

Structure it tightly. Who you are and what you have done, in two lines. The connection to their work, in two lines. What you are asking for, stated plainly: a PhD position from a specific intake, or a funded internship. Under 200 words in total, with a one-page CV attached as PDF, named properly.

Get the register right. Do not write "Respected Sir/Madam" or "kindly do the needful"; use their title and surname, write in ordinary professional English, and do not apologise for writing. Never send a mass email with visible other recipients, and never send the same text to two professors in the same group.

Then be patient and realistic. Send in batches of five or six, keep a spreadsheet, follow up exactly once after two weeks, and treat silence as information rather than rejection. My own success rate was low. It only takes one reply.

Your CV is a curated argument for why a professor should spend their limited time on you. It is not a complete history of your life, and everything on it that does not advance that argument is actively working against it.

The most important entry is your thesis or project work, and it should be written concretely: what you set out to measure, which technique you used, what came out. "Synthesised five-cation perovskite oxides by sol-gel and confirmed single-phase formation by XRD with Rietveld refinement" tells a reader something. "Worked on nanomaterials" tells them nothing.

List only tools you can genuinely use, because you will be asked. A short honest list of MATLAB, Python and Origin with real work behind each is far stronger than twelve logos you once saw in a workshop. The same applies to instruments: distinguish between what you have operated and what you have watched.

Typeset it in LaTeX and keep it to one page as a student. Both signal precision before a single line is read. Cut the school marks, the photograph, the declaration and the hobby section that Indian CV templates insist on; the international convention is leaner and you should follow it.

Start with the research question, not the university ranking. A mid-ranked university with the right group and a professor who supervises properly will do more for you than a famous name where you are one of thirty students and never speak to the advisor.

Find the groups by reading. Take the two or three topics you care about, search Google Scholar for the last three years, and note who keeps appearing as last author. That list is your target list, and it will look nothing like a ranking table.

Then check the things the group webpage will not tell you. How recently were their students publishing, and did those students finish? Is funding active? Does the group work on instruments you would actually get access to, or is the equipment listed because the department owns it?

Finally, write to their current students, not only to the professor. One or two will reply honestly about supervision style, working hours and whether people are treated well. That conversation is the single most useful thing I did, and it is the step most applicants skip.

This blocks more Indian students than it should. The tests are expensive, the slots are scarce, and for many research programs they were never required in the first place.

Many programs in Taiwan, South Korea, Japan and parts of continental Europe waive the English test requirement when your prior degree was taught in English, which is true for almost every Indian science graduate. NTU and NTHU both allow this route.

What you usually need instead is a medium-of-instruction letter from your examination office or registrar, on letterhead, stating that your degree was taught and examined in English. Request it early: Indian university administrations are slower than application deadlines assume, and this one sheet of paper has cost people an intake.

Always confirm the requirement against the specific department page for the specific intake, and if it is ambiguous, email the graduate admissions office and keep the reply. Do not rely on a forum post, or on this page.

Taiwan's Taiwan Experience Education Program is, in my experience, the most accessible fully funded short-term research internship open to Indian students in semiconductors and materials science. It typically covers flight, accommodation and a monthly stipend, so it does not depend on your family's means.

Its real value is not the stipend. A funded internship puts you inside a group for months, which is long enough for a professor to know whether they want you as a doctoral student. My TEEP at NTHU's College of Semiconductor Research in 2025 led directly to my PhD position at NTU, and that is the ordinary way these things happen rather than a lucky exception.

Apply through a professor, not only through the portal. Identify the group first, write the cold email described above, and mention TEEP as the funding mechanism for a visit you have already thought through. A professor who wants you will help the application along.

Within India, the INUP-i2i workshops at the IITs and the CeNSE winter school at IISc are the closest equivalents: real cleanroom and instrument time, national selection, and credibility on a CV. I did both, and they are what made the Taiwan application credible.

Four tools cover most of what a physics or electronics researcher does day to day, and none of them takes a course to start.

MATLAB remains the lingua franca of optical simulation and signal processing, and in metrology groups it is often what the existing code is written in. Python with NumPy and SciPy is essential for anything data-driven, and unavoidable if you go anywhere near machine learning; learn to write it as an engineer, with version control, not as a notebook you never reopen.

Origin is still the standard for scientific plotting in materials science, and being the student whose XRD and UV-Vis figures are publication-ready is a quietly large advantage. LaTeX is non-negotiable for theses and papers, and the earlier it stops being an obstacle, the earlier your writing gets attention instead of your formatting.

Learn them on your own data during your M.Sc., not the week a deadline arrives. The skill that actually matters underneath all four is being able to take a measurement, analyse it, plot it defensibly and write it up without supervision. That is what a supervisor is really buying.

Assume you will not be paying. Doctoral study in the sciences is normally funded, and if a program expects you to fund yourself for years, treat that as a warning rather than an opportunity.

Funding comes in three shapes. Government scholarships, such as Taiwan's MOE and NSTC schemes, which you apply for separately and early. University or departmental fellowships, which often arrive automatically with admission. And project funding from your professor's own grants, which is the most common and the least advertised, and the reason a professor who wants you can sometimes create a position.

Ask about money directly, once you are in a real conversation. The correct question is what the monthly stipend is, whether tuition is covered, and for how many years it is guaranteed. Asking this is normal and no reasonable supervisor will hold it against you.

Also budget for the gap. Flights, initial rent and the first month before a stipend arrives are real costs that fall in the worst possible window. Plan for them explicitly rather than discovering them on arrival.

Work backwards from the intake, not forwards from today. For a September or February start, you want to be shortlisting groups and reading their papers around ten to twelve months out, sending the first emails eight or nine months out, and holding documents (transcripts, medium-of-instruction letter, recommendation letters, CV) ready six months out.

Recommendation letters are the step most likely to slip, because they depend on other people. Ask your thesis supervisor early, give them your CV and a short note on what you are applying for, and remind them politely twice.

Expect this to take more than one cycle, and expect most emails to go unanswered. That is not a verdict on you; established professors receive far more mail than they can answer. The only thing that behaves reliably is volume plus specificity, sustained over months.

If a cycle fails, use the year to strengthen the argument rather than to repeat it. A workshop with genuine instrument time, a paper submitted, a real skill added, and the next round of emails is materially different. That is roughly the path I took, and there was nothing exceptional about it beyond not stopping.