Observability, Data Scientist

At Graphcore, we’re building the future of AI compute. We’re a team of semiconductor, software and AI experts, with deep experience in creating the complete AI compute stack - from silicon and software to infrastructure at datacenter scale. As part of the SoftBank Group, backed by significant long-term investment, we are delivering key technology into the fast-growing SoftBank AI ecosystem. To meet the vast and exciting AI opportunity, Graphcore is expanding its teams around the world. We are bringing together the brightest minds to solve the toughest problems, in a place where everyone has the opportunity to make an impact on the company, our products and the future of artificial intelligence.

Essential:

• BSc/MSc/PhD in Data Science, Computer Science, Statistics, or related field.
• Strong experience with time-series data analysis and large-scale telemetry datasets.
• Proficiency in Python (NumPy, Pandas, SciPy, ML frameworks).
• Experience with:
  • Anomaly detection techniques (statistical + ML-based)
  • Predictive modeling and forecasting
  • Signal processing techniques (filtering, smoothing, FFT or similar is a plus)
  • Data visualization tools (e.g., Grafana, Tableau, Plotly)
• Familiarity with:
  • Time-series databases (e.g., Prometheus, InfluxDB)
  • Observability stacks and monitoring systems
• Strong understanding of data pipelines and distributed systems.
• Excellent communication skills - ability to translate data insights into engineering actions.

Desirable:

• Experience with data center infrastructure, hardware telemetry and cloud platforms.
• Knowledge of monitoring, observability and management solutions in use by hyperscalers.
• Familiarity with root cause analysis in distributed systems.
• Experience deploying models into production.
• Understanding of system architecture (CPU/GPU/accelerators, networking).
• Experience with real-time analytics frameworks (Kafka, Flink, Spark).

Data Analysis & Modelling:

• Analyse large-scale telemetry datasets (metrics, logs, events) across multiple layers: data center (clusters, networking, cooling), system (servers, accelerators), silicon-level (on-chip sensors, performance counters, voltage, thermal, error signals).
• Develop anomaly detection models for infrastructure-level events and hardware/silicon anomalies (e.g., thermal hotspots, voltage instability, error rate drift).
• Build predictive models for fault detection and failure forecasting (e.g., hardware degradation, thermal issues, network anomalies).
• Apply statistical and machine learning techniques to identify patterns and root causes.

Automation & Intelligence:

• Design automated remediation strategies informed by both system and hardware-level signals.
• Collaborate with engineering teams to integrate models into observability and control systems.
• Enable closed-loop optimization using real-time hardware telemetry streams.

Data Engineering Collaboration:

• Work with telemetry engineers on data ingestion pipelines, ensuring data quality and usability.
• Help define schemas, feature extraction pipelines, and aggregation strategies (e.g., down-sampling, windowing).
• Optimize use of time-series databases and analytics platforms.

Visualization & Reporting:

• Build dashboards and visualizations for operational insights (Grafana, Superset, etc.).
• Present complex data in clear, actionable formats for engineering and leadership.
• Define KPIs and health metrics for infrastructure systems.

Cross-Functional Collaboration:

• Partner with platform, hardware, and software teams to understand system behaviour.
• Support debugging, performance analysis, and benchmarking efforts using telemetry data.
• Contribute to reference designs and best practices for observability and analytics.

Applicants for this position must hold the right to work in Poland. Visa sponsorship or support for visa applications is not provided.