Senior MLOps Engineer

What you will do

The Data & ML Ops team belongs to a larger group - Dev Foundations, which is focused on building tools and services for our internal customers within Tractable: researchers, product engineers, ops specialists, etc. We have 4 teams in Dev Foundations tackling different aspects of the space: Infrastructure & Security, Analytics, QA and Data & ML Ops. As a Senior Data Ops Engineer in the Data & ML Ops team, you will be collaborating with peer teams in Dev Foundations to provide a solid technical foundation, with product engineering teams building our asset appraisal platform and with researchers creating & iterating on the models which underpin our products.

We are looking for a Senior MLOps Engineer to build and support systems that enable the core mission of Tractable - to make applied AI possible - by optimising the end-to-end Machine Learning life cycle. The vision of the Data & ML Ops team is to enable researchers to spend 80%+ of their time solving tricky ML problems rather than dealing with engineering/infra/ops challenges. 

You will help mature our ML and Data platform to a world-class state. You will influence the scope and technical direction as well as champion best practices within the team. You have a relentless focus on user experience (researchers, data scientists and product engineers) and you care deeply about what your team is building to make sure it will have the biggest impact on your users. You will be a strong mentor, nurturing and encouraging the supportive environment to enable the team to do their best work.

The role: 

You'll play a key role in developing our ML & Data platform from ground up, as part of a small, high-performing team. You will influence the scope and technical direction as well as champion best practices within the team. You will continuously pursue clean code practices and contribute towards overall platform architecture, collaborating with our other Engineering and Product teams.

You will:

• Work with engineers, researchers and data scientists to build the next generation of Tractable’s ML & Data platform
• Help identify and realise capabilities in our ML & Data platform that massively speed up getting research to production across dataset & model management, model training, model serving, labelling, Data & ML pipeline orchestration and more
• Support Research and Product Engineers with tools and processes to enable a seamless data flywheel
• Deploy and continuously develop robust infrastructure, using best practices for managing infrastructure-as-code
• Solve cost and performance scalability challenges in both model training and model serving
• Run, monitor and maintain business-critical, production systems
• Adopt open-source technologies to best leverage our in-house resources
• Promote engineering best practices throughout the team
• Suggest, collect and synthesise requirements to create an effective feature roadmap

Tech Stack:

We rely heavily on the following tools and technologies, but we are likely to explore new technologies / frameworks as we are building the platform from ground up. You don't need to have prior experience in all of them, and we actively encourage diverse views on what the best tools for the job are. We’re just keen to know that you're willing to break things, fix things, learn fast and help build a great team that is capable of building a platform that delights our customers.

• Main Infrastructure: AWS (EC2, S3, MSK, Lambda, StepFunctions, Glue, IAM, Cognito, Systems Manager, CloudWatch, SQS, Route 53, Sagemaker), Apache Kafka (AWS MSK), Kubernetes, Datadog (Metrics, Logs, Synthetics), Pagerduty
• Main CI/CD: Terraform, Docker, Harness
• Main Databases: Postgres / RDS, Redis, DynamoDB
• Main Languages: Python, Node + Typescript, SQL (Postgres)
• Main Data stack: AWS MSK, AWS Lambda, AWS Redshift, dbt, Airflow, Airbyte, AWS Glue
• Main ML stack:  Triton, TFServing, KServe, AWS Sagemaker, AWS Lambda, AWS MSK, sync/async APIs, Weights & Biases, Tensorflow, Pytorch, dvc, Dagster/Flyte, Streamlit

We encourage you to drop us a line even if you don’t have all the points above. That’s a lot of different areas of responsibility! We will help you pick them up because we believe that great people come from all walks of life.

What you need to be successful:

A strong ML Engineer who is passionate about building platforms that massively reduce lead time from bringing Machine Learning research to production. You have a solid background in core software engineering principles and a good understanding of the difficulties faced by data scientists. A few things we are particularly interested in seeing from you:

• Great communication skills and a collaborative mindset
• An ability to catalyse both process and technical change in a complex, highly cross-functional environment
• 2+ years of experience in building scalable Machine Learning systems
• Have experience building and/or managing scalable data infrastructure (data ingestion, data lake, data warehouse, data orchestration)
• Strong programming experience, from self-contained algorithms to complex object modelling design
• Worked with Python in a professional environment for 2+ years
• Experience working with and scaling model training across GPU clusters
• Experience in building data pipelines and managing data infrastructure
• Experience deploying and managing infrastructure-as-code
• Able to design scalable, robust, fault-tolerant system architecture and compare trade-offs (distributed systems experience a plus)
• Experience building robust, intuitive tooling to support internal users (e.g. common ML libraries, CLIs etc.)
• Numerical computing experience
• Cares about team practices / pairing / advocate of CICD
• Basic ML knowledge, with experience in training computer vision models at scale highly desirable

What’s in it for you

Generous financial reward for your effort

• Yearly compensation reviews
• Generous equity package
• 5% employer match on pension

Time off and flexible working

• 25 days paid annual leave + bank holidays
• Ability to work from abroad for up to 6 weeks/year
• Competitive maternity + paternity leave
• Flexible hours and hybrid working 
• Additional leave to support you when you need it, including sick pay, compassionate leave, or paid time-off to recharge after an intense work period

Support for your health and wellbeing

• Highest tier of private health coverage through Bupa 
• Eyewear and eyetests expense scheme
• Annual Headspace Subscription
• Mental health coaching with Sanctus

Additional perks 

• Workplace Nursery Benefit
• L&D budget to use on Learnerbly (our learning platform)
• WFH expense scheme
• Cycle to Work scheme
• Charity donation scheme with Tyve

#LI-HM1