Giant Leap Trainee, Machine Learning Automation Developer

Vaisala is a global leader in measurement instruments and intelligence helping industries, nations, people, and the planet to thrive. From predicting hurricanes to optimizing renewable energy production, our technology is used where it matters the most - from data centers, windfarms and laboratories to airports, the Arctic and even the surface of Mars. Our team of over 2,300 experts and 59 nationalities around the world is committed to taking every measure for the planet.

Driven by our shared purpose, curiosity, and pioneering spirit, we stay ahead and make a difference. At Vaisala, you don't have to fit in to belong.

Are you ready to take a Giant Leap?We're looking for a Giant Leap Trainee - Machine Learning Automation Developer in our Lousiville, Colorado office. This is a full-time, hybrid, paid internship role.

Giant Leap is a unique opportunity for you to be the project manager of your own project, with the support of your teammates and project supervisor. Every year, these projects are carefully selected by our leadership teams, meaning that you get to work on real‑life questions that are important for us as a company.

We invest in your growth with training sessions throughout the summer. Along with lessons from your own field and work life in general, you have a chance to learn, for instance, presentation skills, project management and problem solving. Connections built with fellow Giant Leapers, Vaisala's brilliant experts and our leadership form an invaluable network for your career. Many of our former Giant Leapers have also gone on to build impressive careers at Vaisala after the program.

We're building a production‑grade Weather Sensor Exception Classifier service that turns Nagios critical alerts into actionable, root‑cause‑informed troubleshooting steps in seconds to minutes, and this role takes it the rest of the way: reliable, scalable, and ready to operate across multiple observation networks.

• Productify and extend an existing proof‑of‑concept outage/exception classifier into a robust & extendable production service.
• Refine a pipeline that ingests critical‑state alerts from monitoring services, enriches them with xWeather data & internal telemetry, and outputs:
  • likely root cause
  • confounding factors (weather, network conditions, sensor patterns)
  • recommended troubleshooting / recovery steps
• Develop and maintain a full‑stack LAMP implementation supporting the ensemble service gestalt (API layer, backend, persistence, and UI/ops views as needed).
• Implement deterministic ML / rules + trend‑based inference using historical time series and known failure modes.
• Integrate results into operational workflows to support rapid triage and reduce context‑switching for NOC operators (supporting "0‑click tickets" style automation).
• Design for scale across networks such as lightning, Wind Cube lidar, weather radar, airports, and other observation systems.
• Collaborate closely with NOC stakeholders to validate outputs, tune inference logic, and ensure the service improves real‑world resolution time.

• Strong experience building full‑stack LAMP services (Linux/Apache/Maria
  
  DB/Mongo
  
  DB/Python).
• Solid Python REST microservice development experience;
  FastAPI / Flask.
• Experience working with time‑series data (ingestion, normalization, trend detection, anomaly/exception patterns).
• Practical approach to "deterministic machine learning" / explainable inference (rules, heuristics, feature‑based classification, or lightweight models where behavior is understandable and supportable).
• Ability to ship production software: testing, logging/observability, reliability, clear APIs, and attention to comprehensive documentation.
• Comfort collaborating with operations teams and iterating quickly based on feedback from diverse sources.

• Critical alerts are automatically enriched and translated into high‑confidence, explainable recommendations quickly enough to support incident response in real time.
• The service scales beyond the initial network and becomes a reusable pattern for other monitored systems.
• NOC operators spend less time on repetitive…