AI Scientist, Platform
AvathonPleasanton, California, United StatesOn-Site
Needs a free Superjobs account. You'll finish on Avathon's site.
Skill score
Summary
The AI Scientist builds shared machine learning capabilities for the Avathon platform, including normal behavior models, anomaly detection, knowledge graph methods, and AI agents, and works with platform engineering to move models from research into production. It suits someone with strong ML depth and experience shipping models used across many products or customers.
- Function
- Data Science
- Estimated salary
- Est. $110k–166k/yr
- Estimated for a mid-level role, ~5 yrs experience, United States. Not published by the employer.
- Work mode and location
- On-Site, Pleasanton, California, United States
- Experience
- 5+ yrs
Job description
Overview
We are hiring an AI Scientist to build the shared AI capabilities of the Avathon platform. Our applications in planning, logistics, PLM, asset performance and machine vision all rely on the same core: Normal Behavior Models that learn what normal looks like, a Computational Knowledge Graph that links industrial data, a Model Hub that serves ML and LLM models, and AI agents that reason over all of it. You will improve that core so every application gets better at once. This role sits in Product Engineering. You will work with platform engineering to take models from research to production.
Roles and responsibilities
Core Models
- Improve Normal Behavior Models and anomaly detection for time series and event data across industries
- Build methods for data quality and gap filling on messy industrial sensor and business data
- Evaluate and adapt time series and multimodal foundation models for industrial use
Knowledge Graph and Agents
- Build methods for ontology discovery, entity resolution and mapping raw data into the knowledge graph
- Build and evaluate AI agents that query the graph, call models and recommend actions, with clear guardrails and human approval
- Design evaluation frameworks so we know when an agent or model is good enough to ship
Platform and Productization
- Work with the Model Hub team on model serving, monitoring, retraining and cost
- Package methods as reusable platform services that application teams can use without starting from scratch
- Write up methods and results clearly for product, customers and leadership
Skills required
Top skills
- Machine learning
- Time series anomaly detection
- Knowledge graphs
- LLMs and agents
All skills (8)
- Python
- PyTorch
- Hugging Face
- Model serving and monitoring
- Normal behavior models
- Foundation models
- Computer vision
- Entity resolution
Qualifications
- PhD in Computer Science, Machine Learning, Statistics or a related field, or an MS with 5+ years of relevant industry experience
- Strong foundation in machine learning, with depth in at least two of: time series and anomaly detection, knowledge graphs and graph learning, LLMs and agents
- Experience building ML or AI capabilities used across many products or customers
Nice to have
- Experience at Palantir, C3 AI, Databricks, Google, AWS, Microsoft or an industrial AI company is a strong plus
- Strong Python and the standard ML stack (PyTorch, Hugging Face, LLM frameworks)
- Experience shipping models to production, not only notebooks
- Publications or patents in related areas are a plus
Other details
Pay Range: $160k - $240k salary annually. Pay for this position is based on a number of factors including geographic location and may vary depending on job-related knowledge, skills, and experience.
Location: This role is not remote. Candidates must be based in the Bay Area, CA and are expected to report to our Pleasanton office 5 days a week.
Company info
About Avathon
Avathon is the leading Industrial AI autonomy platform, helping customers across heavy industries -- energy, mining, manufacturing, aerospace, defense, and logistics -- accelerate the journey toward autonomous operations. Our platform is built on a Computational Knowledge Graph foundation that contextualizes and connects operational data across siloed systems, bringing together time