Intern AI Researcher: Anomaly Detection for Predictive Maintenance in Distributed Storage
Posted Updated
Description
About Scality
Scality builds data infrastructure software for enterprise AI, cyber resilience, and sovereign control at multi-petabyte to exabyte scale. Its AI Data Infrastructure platform reduces operational burden through AI-powered autonomous operations while aligning the right storage media, performance, and protection to each workload through human-approved lifecycle policies.
Built on CORE5 cyber resilience and open-code principles, Scality software delivers extreme performance, operational simplicity, and sustainable economics.
The world's most demanding enterprises and government organizations depend on Scality to power AI initiatives, defend critical data, and build infrastructure that lasts for decades. Recognized as a leader by Gartner, Scality is where AI remembers, learns, and thinks. F
Follow us on LinkedIn. Visit and our blog.
Context & Motivation
A distributed storage cluster continuously emits telemetry at every level, from the platform and S3 service down to each server and disk. A failing disk, server or storage node leaves traces in these multivariate time series, from subtle shifts in a few metrics to system-wide disruption. Detecting and diagnosing these anomalies early is critical for:
- Predictive maintenance: isolating or replacing a component before it impacts service
- Faster diagnosis: pointing support engineers to the root cause, not just to an alert
- Higher availability and durability of customer data
This internship is part of a joint research project between Scality and an INRIA research team. The project released SHAD (Scality High-dimensional Anomaly Detection), an open benchmark published at NeurIPS 2026 (Evaluations & Datasets Track). It contains 215 real 171-dimensional time series from Scality RING clusters, annotated for injected disk, server and storage-node failures. SHAD's baselines show the problem is far from solved. No detector handles every failure type, per-dimension attributions find root causes no better than chance, and off-the-shelf LLMs cannot interpret the anomalies. This internship builds on SHAD to close these gaps.
More details on the project:
Your Mission
As part of our R&D team and in close collaboration with INRIA researchers, you will develop AI models that detect, explain and interpret anomalies in storage-cluster telemetry. Your work will include:
1. State of the Art
Review the literature and open-source projects on multivariate time series anomaly detection, root-cause attribution and LLM-based interpretation of time series. Also survey recent work on AI agents that automate research and experimentation. Starting points include SHAD and the TSB-AD benchmark.
2. Anomaly Detection & Diagnosis Models
Design, train and benchmark models that improve on SHAD's baselines at each stage of the pipeline:
- Detection that stays accurate across disk, server and storage-node failures (e.g., ensembles, hybrid models, time series foundation models)
- Explanation: attributing each anomaly to the dimensions that actually cause it
- Interpretation: turning detections into actionable diagnoses (failure type, criticality, affected component)
- Early detection of progressive degradations, the core of predictive maintenance
3. Fleets of AI Agents for Research
Set up and manage fleets of AI agents that run your studies at scale:
- Design agent workflows that launch experiments, run benchmarks and compare methods
- Orchestrate and monitor many agents working in parallel
- Build safeguards for rigor: experiment tracking, reproducibility, independent verification of results
- Assess what agent-driven research changes in practice: speed, coverage, failure modes
4. Evaluation & Validation
Evaluate every model with rigorous, reproducible protocols. Use them to:
- Measure each stage with SHAD's metrics (VUS-PR for detection, NDCG for explanation, a questionnaire for interpretation)
Quantify performance per failure family and scenario - Identify which approaches are ready to transfer to production monitoring
Expected Outcomes & Valorization
Depending on results, the work may be valorized through:
- A scientific publication with the Scality and INRIA researchers
- New methods and baselines contributed to the open-source SHAD benchmark
- Reusable agent-based research workflows for the R&D team
Technical Stack
- Python (primary language for modeling, experimentation and analysis)
- PyTorch and scikit-learn (deep learning and classical models)
- SHAD benchmark (github.com/scality/shad) and TSB-AD (github.com/TheDatumOrg/TSB-AD) for data, detectors and evaluation
- LLM-based AI agents and their orchestration tooling
Candidate Profile
- Final-year student in a Master's program or Engineering school (Bac+5), in machine learning, data science or computer science
- Solid foundations in machine learning; experience with time series or anomaly detection is a strong plus
- Strong interest in LLMs and AI agents, and curiosity for new ways of doing research
- Good Python skills and hands-on experience with a deep learning framework such as PyTorch
- Rigorous and critical with experimental results, especially when agents produce them
- Autonomous and curious, able to communicate results clearly in written and spoken English
- Interest in distributed systems or storage is a plus, not a prerequisite
Work Environment
You will join a fast-paced, lean AI R&D team. French and English is our primary working language. You will benefit from:
- Mentorship from R&D engineers
- Close collaboration with INRIA researchers working on time series anomaly detection
- Access to real telemetry from Scality RING clusters and to the AI agent tooling the team uses for its studies
- A high-trust environment where your findings will directly influence engineering decisions
Why Join Scality?
- Work on a concrete research problem with real-world industrial impact
- Build on a benchmark published at NeurIPS and contribute to open research
- Practice a new way of doing research, with fleets of AI agents as research assistants
- Be part of a team building infrastructure trusted by Fortune 500 companies
- Potential to publish research or continue as a full-time engineer after the internship
Skills
As published by lever
Resume/CV, Full name, Email, Phone, Current location, Current company, LinkedIn URL, Twitter URL, GitHub URL, Portfolio URL, Other website
Sign in to read and join the discussion.