AI Engineer, Intern

Postman Postman · Enterprise · Berkeley, CA · AI

AI Engineer Intern role focused on building and validating AI models, preparing data pipelines, and assisting with productionization tasks like inference APIs and model optimization. The role involves working with large-scale AI and Agentic systems from data to deployment, with a focus on learning and contributing to real systems under mentorship.

What you'd actually do

  1. Prepare data pipeline and AI store design under senior engineer guidance.
  2. Build and validate AI models using PyTorch or JAX(Optax/Orbax / TensorStore/Grain) and similar tools.
  3. Run large-scale experiments, evaluate models against defined metrics, and support ablation studies and error analysis.
  4. Help build inference APIs and batch scoring workflows; integrate AI outputs with backend services.
  5. Support AIOps practices already in place on the team: model versioning, CI/CD pipelines, monitoring dashboards, and logging.

Skills

Required

  • BS/MS/PhD in Computer Science, Data Science, or a related field
  • Foundational experience in AI/ML systems
  • Python/Rust fundamentals
  • PyTorch or JAX
  • Pandas
  • SQL
  • Clear written and verbal communication

Nice to have

  • familiarity with a cloud platform (AWS/GCP/Azure)
  • Docker
  • model-serving tool (vLLM, Triton, Ray Serve)

What the JD emphasized

  • own discrete pieces of real systems
  • learning objective, not an expected independent deliverable
  • you're not expected to resolve these alone, but surfacing them is part of the job

Other signals

  • AI Engineer Intern
  • large-scale AI and Agentic systems
  • data pipeline to production deployment
  • own discrete pieces of real systems
  • Build and validate AI models
  • Run large-scale experiments
  • evaluate models against defined metrics
  • support ablation studies and error analysis
  • Help build inference APIs
  • batch scoring workflows
  • integrate AI outputs with backend services
  • Support AIOps practices
  • model versioning
  • CI/CD pipelines
  • monitoring dashboards
  • logging
  • Assist with model optimization work
  • quantization
  • batching
  • distillation
  • Contribute to root-cause analysis on production issues
  • Document experiments
  • design decisions
  • runbooks
  • Flag fairness
  • interpretability
  • privacy concerns