Research Engineer / Research Scientis, Ideogram

Post-training

Salary not provided

+ Equity

Docker

Kubernetes

Tensorflow

CUDA

PyTorch

Senior and Expert level

New York

Toronto

Generative media tools

Open for applications

Generative media tools

21-100 employees

B2CB2BArtificial IntelligenceArtMedia

Open for applications

Salary not provided

+ Equity

Docker

Kubernetes

Tensorflow

CUDA

PyTorch

Senior and Expert level

New York

Toronto

21-100 employees

B2CB2BArtificial IntelligenceArtMedia

Company mission

To make professional-quality design accessible to everyone.

Job

Company

Role

Who you are

5+ years of experience in developing machine learning models in JAX, PyTorch, or TensorFlow
Experience in implementing Machine Learning foundations (e.g., Transformer, VAE, Denoising Diffusion models) from scratch
Track record in machine learning innovation and familiarity with Deep Learning and advanced Machine Learning
End-to-end understanding of generative media applications and excitement for pushing the state-of-the-art in generative AI
Ability to debug machine learning models to iteratively improve model quality and performance
Nice to have: Familiarity with kubernetes and docker
Optional: Experience in low-level machine learning optimization, e.g., writing CUDA kernel code

What the job involves

In this role, you will develop the post-training pipeline for our text-to-image foundation models end to end, from data strategy to deployment, advancing techniques such as RLHF, RLAIF, and work on personalization/customization
You will contribute to post-training research that drives measurable gains, and implement and maintain high-throughput fine-tune/eval pipelines
You'll work with a creative and ambitious team of engineers and researchers who are building the future of the creative economy

Our take

With text-to-image software well and truly in the public conscience, Ideogram has launched a platform to support creativity, while capitalising on the popularity of AI image creation services.

Its generative AI platform requires a prompt and rendering style (think anime, fantasy, or cinematic) from which it creates a high quality image for the user. As with other platforms of its type, the images tend to be slightly flawed but interesting, and it appears that almost anything can be created.

Of course, there is a downside to launching into a market that is under such a strong spotlight, especially as many artists raise concerns over their work being used to train image creation AI without their consent. For its part, Ideogram claims to operate with a "high standard for trust and safety", utilizing a $22.3M CAD funding round to grow its team and footprint.

Freddie

Company Specialist at Welcome to the Jungle

Company

Funding (2 rounds)

Feb 2024

$80m

SERIES A

Aug 2023

$16.5m

SEED

Total funding: $96.5m

Company HQ

Financial District, Toronto, ON

Leadership

Mohammad Norouzi

(Co-Founder & CEO)

Previously Senior Staff Research Scientist at Google for 7 years.

William Chan

(Co-Founder & CTO)

Spent 7 years at Google, ultimately as a Google Brain Staff Research Scientist.

Chitwan Saharia

(Co-Founder)

Spent 3 years at Google, ultimately as a Senior Research Scientist. Prior to this, was a Research Assistant at the Montreal Institute for Learning Algorithms.

Jonathan Ho

(Co-Founder)

Previously a Research Scientist for Google and for OpenAI.

Salary benchmarks

We don't have enough data yet to provide salary benchmarks for this role.

Submit your salary to help other candidates with crowdsourced salary estimates.

Share this job

View 10 more jobs at Ideogram