Senior ML Engineer - Real-Time AI Security & Detection

Refractal Ltd; Greater London; Hybrid; GBP 85,000 - 145,000; Full time

Benefits offered by this job

Salary £85,000–£145,000

Competitive equity to share in Refract

Pension contribution

Private medical and dental cover

Visa and relocation support

Job summary

Refractal Ltd in London is seeking a seasoned ML Engineer to own the production models behind detections and the analyses that show performance for each customer. You will train, tune and ship classifiers for prompt injection, data exfiltration and other risky AI activity, and design low-latency models.

The role emphasizes evaluating drift, false positives and latency, building benchmarks, and turning policies into measurable controls.

Qualifications

• 5+ years in applied ML or data science, with models shipped to production.
• Strong Python, PyTorch or similar, and solid statistics for evaluation.
• Hands‑on NLP and LLM work: fine‑tuning, embeddings, classification and LLM-as-judge.
• Experience measuring fuzzy problems, including labelling, agreement and drift.
• Clear communication of trade‑offs to engineers and customers.

Responsibilities

• Train, tune and ship classifiers for prompt injection, data exfiltration, policy violations and other risky AI activity.
• Design low-latency models, including distilled and small models and rules-plus-ML ensembles.
• Analyse production telemetry for drift, blind spots, false-positive cost and new attack patterns.
• Build evaluations and benchmarks for our detection models, and work with security engineers to turn new attacks into training data.
• Turn company policies and regulatory requirements into measurable controls.

Job description

Refractal Ltd in London is seeking a seasoned ML Engineer to own the production models behind detections and the analyses that show performance for each customer. You will train, tune and ship classifiers for prompt injection, data exfiltration and other risky AI activity, and design low-latency models.

The role emphasizes evaluating drift, false positives and latency, building benchmarks, and turning policies into measurable controls.