Senior Machine Learning Engineer, Grafana Labs

• You may not meet every requirement, and that’s okay. If this role excites you, we’d love you to raise your hand for what could be a truly career-defining opportunity
• We know it is rare to find everything. Strong candidates should demonstrate credible ability across all three core areas below and be particularly strong in at least two
• Recommendation and personalization science: you have built recommendation, ranking, search, matching, propensity, or next-best-action systems. You are comfortable beginning with simple, explainable approaches when they are the best way to learn
• HTTP/gRPC, streaming, Go/TypeScript previous experience in distributed systems
• Applied model ownership. You have personally built, validated, monitored, and iterated on models used in a product or operational environment. You can work effectively in version-controlled codebases and collaborate with engineers on production implementation
• You should also be a strong product thinker and technical communicator. You can take an ambitious and ambiguous objective, identify the most important unknowns, and create a sequence of models and experiments that steadily improves the product
• Experience with content, education, onboarding, or learning recommendation systems

Desirable

• The long-term vision is ambitious, but we do not expect it to arrive in one release.
• We are looking for someone who can understand the whole problem, establish strong foundations, and ship measurable improvements into the existing recommender one iteration at a time
• Evolve the Interactive Learning Plugin's recommendation system
• Develop increasingly personalized approaches to candidate selection, ranking, sequencing, and next-best-action recommendations
• You’ll own a real-time recommendation service
• Build and operate applied models
• Develop, validate, version, monitor, and iterate on models used by the recommendation system
• You’ll own model training & serving
• Define what recommendation quality means
• Develop offline, online, and longitudinal measures of recommendation performance
• You’ll own feature pipelines, monitoring of the model and architecture
• Ship incremental improvements
• Use the data and infrastructure available today while identifying the instrumentation and platform capabilities needed tomorrow
• Integrate improvements into the existing recommender rather than waiting for a complete replacement system
• Partner across disciplines
• Work closely with software engineers & data analysts to productionize models and integrate them safely into the recommender service
• Partner with the Product Analytics team on metric definitions, instrumentation, data quality, dashboards, and experiment analysis
• Collaborate with Developer Advocacy, Docs, Product, Engineering, GTM, and other teams to translate ambiguous needs into testable hypotheses and measurable product decisions
• Explain modeling choices, tradeoffs, uncertainty, and results clearly to both technical and non-technical audiences