Responsibilities
- Design and improve ML/DL algorithms on large-scale data to automate test and triage workflows.
- Build and run agentic LLM systems (e.g., Claude or Gemini) that automate triage, from prototype to production.
- Run multiple production pipelines: monitor them, respond to issues, and fix the root causes.
- Work with stakeholders across data science, ML, autonomous drive planning, and quality assurance.
- Build monitoring tools and use them to keep improving your algorithms in production.
Requirements
- PhD or Master's in a STEM field and 4+ years in Triage, Performance Analytics, for fortune 500 including infrastructure as code, ML, NLP, or deep learning.
- Strong productionization, CI/CD experience for medium to large projects Python,, with a solid grasp of dashboards, eval, algorithms and data manipulation.
- Experience with RAG, PyTorch, TensorFlow, or HuggingFace
- Has built or managed agentic LLM systems (Claude or Gemini) in a professional setting with large dataset to satisfy multiple stakeholders.
- Has run multiple production pipelines on platforms like Databricks and AWS, including fixing production issues.
- Experience with self correcting models, installation and maintenance of LLMs, data visualization tools like Looker/Dash/Databricks Dashboards
- Experience in Performance Monitoring at Scale, Support environments, autonomous driving, robotics.
- Strong verbal and written communication skills
As a Software Development Engineer in Test on the Product Software Quality Assurance team, your role will focus on building scalable integration automation frameworks and enhancing existing frameworks for automated testing of mobile applications. You will design and implement functional and integration automation tests for mobile app used by the customer fleet, ensuring comprehensive test coverage and reducing manual testing efforts.
By prioritizing automation for repetitive tasks, developing robust regression suites, and focusing on edge cases and high-stress scenarios, you will ensure the reliability of safety-critical tools.