Software Engineer (Machine Learning)

About Netcraft

The role

Day-to-day work

• Build, evaluate, and tune classifiers used to block and take down content.
• Deploy and operate Kubernetes model-serving infrastructure, including LLM inference.
• Optimize models for high-volume, low-cost serving.
• Experiment against live data, measure precision, recall, and false-positive rates, and decide what ships based on results.
• Integrate models into production classification platforms.
• Build feature-extraction, scheduled-retraining, experiment-tracking, and monitoring pipelines.
• Investigate production model regressions and performance bottlenecks.
• Work with engineers and analysts who depend on the classifiers.

What you'll need

• Strong web fundamentals, including HTTP, DNS, and how websites work from request to rendered page.
• Working knowledge of machine learning training, evaluation, overfitting, and precision/recall trade-offs, gained through industry, research, or substantial personal projects.
• Clear written and verbal communication, an empirical mindset, attention to detail, and a strong sense of responsibility.
• Willingness to learn Go and modern Perl, including asynchronous programming.

Bonus experience

• Interest in cybercrime or cybersecurity; model fine-tuning and evaluation; LLM agents, evaluation, or serving such as vLLM or llama.cpp.
• Familiarity with Kubernetes, Flux, Argo, or API gateways; multimodal models, embeddings, or similarity search at scale.
• MLflow, Argo Workflows, SQL databases, and query optimization.

Location preference

• London is preferred; Manchester and Bath are also options.

Reward package

• Benefits include 33 days of annual leave including public holidays, enhanced family leave and pay, enhanced sick pay, matched pension contributions up to 5% plus National Insurance savings, private health cover, and life assurance.
• An equity tracking scheme is available subject to eligibility criteria.