Skip to contentSkip to footer

Applied Scientist Interview Questions

Salaries
Interviews

1,159 applied scientist interview questions shared by candidates

You'll be given a GitHub repo with a project to do. The task requires the training of an LLM and some specific tasks, such as writing a custom cost function to take into account data imbalance and a large cardinality of slots with low volume/representation in the training set. To give you an idea, the samples in the training dataset would look like: INPUT: "find me the nearest gas station off of 5th ave north." OUTPUT: "[IN:GET_LOCATION find me the [SL:LOCATION_MODIFIER nearest ] [SL:CATEGORY_LOCATION gas station ] [SL:LOCATION_MODIFIER [IN:GET_LOCATION [SL:SEARCH_RADIUS off of ] [SL:LOCATION 5th ave north ] ] ] . ]" INPUT: "What time do I need to leave Norman to arrive in Oklahoma City by five o'clock?" OUTPUT: "[IN:GET_ESTIMATED_DEPARTURE What time do I need to leave [SL:SOURCE Norman ] to arrive in [SL:DESTINATION Oklahoma City ] [SL:DATE_TIME_ARRIVAL by five o'clock ] ? ]" and so on. The output needs to be machine readable and written in a formal tree-like grammar, but you need to use an unconstrained generative LLM to achieve that.
avatar

Applied Scientist

Interviewed at Unlikely AI

4.6★
Sep 17, 2023

You'll be given a GitHub repo with a project to do. The task requires the training of an LLM and some specific tasks, such as writing a custom cost function to take into account data imbalance and a large cardinality of slots with low volume/representation in the training set. To give you an idea, the samples in the training dataset would look like: INPUT: "find me the nearest gas station off of 5th ave north." OUTPUT: "[IN:GET_LOCATION find me the [SL:LOCATION_MODIFIER nearest ] [SL:CATEGORY_LOCATION gas station ] [SL:LOCATION_MODIFIER [IN:GET_LOCATION [SL:SEARCH_RADIUS off of ] [SL:LOCATION 5th ave north ] ] ] . ]" INPUT: "What time do I need to leave Norman to arrive in Oklahoma City by five o'clock?" OUTPUT: "[IN:GET_ESTIMATED_DEPARTURE What time do I need to leave [SL:SOURCE Norman ] to arrive in [SL:DESTINATION Oklahoma City ] [SL:DATE_TIME_ARRIVAL by five o'clock ] ? ]" and so on. The output needs to be machine readable and written in a formal tree-like grammar, but you need to use an unconstrained generative LLM to achieve that.

Estimate the mass of the earth System debugging (simplified conceptual model of their breathalyzer) test the efficiency of this pump (intentionally bad equipment provided, measure flow rates, voltages, currents, calcuate efficiency in excel).
avatar

Applied Scientist/Systems Engineer

Interviewed at Owlstone Medical

3.8★
Jun 29, 2017

Estimate the mass of the earth System debugging (simplified conceptual model of their breathalyzer) test the efficiency of this pump (intentionally bad equipment provided, measure flow rates, voltages, currents, calcuate efficiency in excel).

Viewing 151 - 160 interview questions

Glassdoor has 1,159 interview questions and reports from Applied scientist interviews. Prepare for your interview. Get hired. Love your job.