ML Data & Platform Engineer | Hybrid & Data Pipelines

Speechmatics; Greater London; Hybrid; GBP 90,000 - 120,000; Full time

Benefits offered by this job

Private Medical

Dental for you and family

Generous holiday allowance

Pension/matching

Job summary

Speechmatics is seeking an ML Data & Platform Engineer to own the infrastructure powering our speech AI models, spanning data pipelines, training, evaluation, and serving. You’ll work across data infra, distributed systems, and production ML, taking ownership of problems end to end.

This role involves building scalable data pipelines, web scraping, and optimistic reliability enhancements, with a hybrid work setup and focus on improving MLOps practices as we scale.

Qualifications

• Strong proficiency in Python and SQL with backend/data engineering experience.
• Hands-on containerisation and orchestration (Docker, Kubernetes) with cloud experience.
• Experience building data pipelines and ETL/ELT at scale, incl. web scraping or automated data collection.
• Understanding ML lifecycle from data to training, evaluation, and serving; observability.

Responsibilities

• Design, build, and maintain scalable data pipelines for training data.
• Develop web scraping and data acquisition solutions at scale.
• Build and operate infrastructure for deploying/evaluating models in production.
• Optimise infrastructure for iteration speed and production reliability.
• Implement observability across data pipelines and ML systems.
• Troubleshoot issues across distributed systems (data, training, inference).
• Improve data and MLOps practices and shape the platform roadmap.

Skills

Python

SQL

Data pipelines

ETL/ELT

Docker

Kubernetes

Cloud

Data quality

Observability

MLOps

Distributed systems

Self-starter

Tools

Docker

Kubernetes

AWS

Job description

Speechmatics is seeking an ML Data & Platform Engineer to own the infrastructure powering our speech AI models, spanning data pipelines, training, evaluation, and serving. You’ll work across data infra, distributed systems, and production ML, taking ownership of problems end to end.

This role involves building scalable data pipelines, web scraping, and optimistic reliability enhancements, with a hybrid work setup and focus on improving MLOps practices as we scale.