Every Mechanical engineering student in India studies Thermodynamics, Heat Transfer, Fluid Mechanics and Refrigeration & Air-Conditioning. Almost none of them know that these four courses are exactly the skill stack the AI-datacenter boom is desperate for. A single NVIDIA GB200-class rack dissipates over 100 kW — more heat per square metre than almost any machine a fresh graduate has ever analysed — and India is building datacenter capacity from roughly 1.4 GW today toward 4 GW+ by 2030. Every one of those megawatts of compute is a megawatt of heat, and someone has to move it: through cold plates, coolant loops, air handlers, chillers and cooling towers. That someone is a mechanical engineer, and there are far too few who can do it.
This internship takes the theory you already passed exams on and makes you apply it to real datacenter problems. You start with the heat-load mathematics of actual GPU server spec sheets, size airflow with psychrometrics against the ASHRAE TC9.9 thermal envelope, and work through room-level cooling: CRAC/CRAH selection, hot/cold-aisle containment, and CFD simulation of airflow so you can see recirculation and bypass instead of guessing at them. Then you go where the industry is going — direct-to-chip liquid cooling, coolant distribution units, rear-door heat exchangers and immersion cooling for 30–50 kW/rack AI rows — and out to the plant: chillers, cooling towers, pumping schemes, free cooling for Indian climates, PUE optimisation and N+1 redundancy design. Every week produces a calculation workbook, a simulation, or a design document that a hiring manager can actually evaluate.
The internship is built for the Indian academic calendar and the AICTE/NEP internship mandate. Take it as a 4-week winter sprint, an 8-week summer internship, or a 6-month final-semester capstone that maps to your project/internship credits. Every track ends the same way: a graded capstone — a complete rack-to-room cooling design for a high-density AI rack row, defended live before a mentor panel — a portfolio of engineering calculations and CFD evidence, an RKR completion certificate, and, for the strongest interns, a referral into the mission-critical MEP and datacenter-operator hiring pipeline.