Associate Data Analyst

1.Personal Details

We'll need these details in order to be able to contact you.

Apply with LinkedIn

First Name

Last Name

Email Address

Phone

LinkedIn URL optional

2. Profile

Résumé / CV

Accepted file formats are .pdf and .docx

3.Questions

Where do you intend to primarily work from?

We can set up a contract in many countries if you have valid local work authorization.

Are you legally authorized to work for us in the country you have stated above?

If hired, we will need to verify your eligibility (your legal right to work in a specific country).

Yes

What is your earliest available start date?

Please provide an approximate start date of when you might be able to join. Knowing when you would be able to join us helps us to update our hiring timelines.

Please take into account any professional, personal, or academic commitments that you would need to fulfil before joining us, independently of when you may receive an offer.

How many hours per week would you ideally like to work on this role?

You may submit an approximate single number or a range.

How did you hear about us or this position?

If applicable, please also share the name(s) of anyone who's invited you to apply below.

Are you happy for us to share your information with related groups?

We're often asked by other people or organizations doing high-impact work for referrals for roles they're hiring for, requests for proposals/requests for grant applications they’re running, or collaborators on projects. If you check "yes" we may proactively share some basic information (your full name, email, resume, how far you made it in our hiring process, and a high-level sentence or two). However, we will not share skills assessment submissions or any scores and will not give more detailed evaluative feedback without checking with you first. Note that if you don't answer this question or answer "no," we will never share your information. And, if you change your mind at any time, just let us know!

Yes

We have a short hiring process for this role, involving only the following screening questions and a final interview. Please don't spend more than 2 hours in answering the questions below (we expect you will need to spend much less than that).

The Besiroglu Neural Machine was trained on 1e6 examples for 10 epochs. The training took 4h on a cluster of 16 GPUs, model Z404. Each Z404 GPU has a peak performance of 2 petaFLOPS. During training, the developers reported a 0.3 utilisation rate. What is the amount of compute, in floating-point operations, that was used to train the system?

Please show your work, and write the result in scientific notation with two significant digits.

For the paper "Xception: Deep Learning with Depthwise Separable Convolutions" (https://arxiv.org/abs/1610.02357), answer the following questions about the paper and the model introduced by the authors:

1. How many times has the publication been cited?
2. How many trainable parameters does the model have?
3. How much data was used to train the model? Report the number of datapoints, as defined here: https://docs.google.com/document/d/1XWLyMzcVfDv4eFQX3yPgM8MZ3_Q1phtIFz9GKv4_KaM/
4. What hardware was used for training? Indicate the model of processors or GPUs.
5. How much compute was used to train the model? Report your answer as a number of floating-point operations.

Explain your reasoning behind each estimate. This might consist of quoting a source that provides the answer, or showing a calculation you did. You may use SI prefixes or scientific notation.

For the paper "Towards a Human-like Open-Domain Chatbot" (https://arxiv.org/abs/2001.09977), answer the following questions about the paper and the model introduced by the authors:

For each of the following benchmarks, what is the current state-of-the-art performance, and which machine learning system has the record for best performance?

a. MMLU
b. ImageNet

Diversity and Inclusion

In an effort to promote equal opportunities and reinforce inclusive hiring practices at our organization we have included below some optional demographic questions. Your responses, or your choice not to respond, is entirely anonymous and will not be associated with your application.

Gender

Ethnicity

Preferred Pronouns

4. Submit Application

In order to contact you with future jobs that you may be interested in, we need to store your personal data.

If you are happy for us to do so please click the checkbox below.

You can view our privacy notice for more information.

Allow us to process your personal information.