Calibration intervals you set yourself, the consumables list that has to be written at handover, which spares to hold and which to leave abroad, and when an annual maintenance contract is worth paying for.
The commissioning engineer leaves, the inauguration photograph is taken, and the project is closed. Two years later half the rigs are unusable and nobody can say when they stopped working. This is not an equipment problem. It is that four separate jobs were never given to anybody.
Write these into a document at handover, with names rather than designations, and review it every year.
Train two people for the technician role, not one. The one you train will be promoted, transferred or will leave, and a laboratory whose only competent operator has gone stops within a term. Insist that training happens on site with the installed equipment after commissioning, and ask for it to be repeated at around six months, when the staff have real questions.
Not everything in a teaching laboratory needs calibrating, and departments waste money pretending otherwise. The test is simple: does this instrument produce a number that a student uses in a calculation, or that the department might put into a report for someone else? If yes, it needs calibration and a record. If it only demonstrates a phenomenon, it does not.
You set the interval, not the supplier. Base it on how heavily the item is used, how badly a wrong reading would matter, and what the drift history shows. A force testing machine used every week by students learning to load it is a different case from one used twice a semester by staff. Write the interval down with the reasoning beside it, because that written justification is what an accreditation panel actually wants to see.
Between calibrations, do intermediate checks and record them: a proving ring or known mass on a testing machine, a certified reference block on a hardness tester, a buffer solution on a pH meter, an ice point on a thermocouple channel. These checks cost minutes and they catch a drifting instrument long before the annual certificate does.
The annual consumables list has to be written at handover, by the commissioning engineer, with part numbers and quantities based on the class size the department actually teaches. If it is not written then, it gets written two years later by somebody guessing, and the guess is always low.
Spares follow a counter-intuitive rule: hold the cheap items and not the expensive ones. Seals, belts, lamps, fuses, thermocouples, filter elements, patch leads and small sensors cost very little and have import lead times measured in weeks once shipping and customs clearance are added. A sealing ring worth a few hundred taka can take a rig out of service for six weeks. Expensive assemblies fail rarely, and money tied up in one on a shelf is money not spent on a second rig. Ask for a recommended two year spares list at bid stage: it is one of the few ways to compare the real running cost of two bids that look alike on price.
Sometimes, and the answer depends on three things: whether the equipment is computer controlled, whether its failure stops a course, and whether the department has a competent technician of its own.
A contract earns its cost on computer-controlled rigs, pilot plants, refrigeration systems, anything with software and a licence, and anything whose failure blocks an assessed course. It rarely earns its cost on cased mechanics kits, hand-loaded strength of materials benches or basic electronics modules.
Where you do take one, insist that the document says five things plainly: how many preventive visits per year and in which months; the response time for a breakdown call, stated in working days; exactly what is included and what is charged separately, particularly calibration, consumables and travel; whether software updates and licence renewals are covered; and what happens when a part is not available in Bangladesh, including who bears the shipping and clearance cost and how long the supplier commits to. A contract that is silent on that last point is silent on the only thing that ever really goes wrong.
None of this survives without a budget line. A department with no recurring maintenance and consumables code has no way to spend money on a laboratory even when it wants to, and the equipment degrades in a way that looks like neglect but is actually an accounting problem. Establish the code in the same year the equipment is bought, while the project still has attention.