About the company
MongoDB is a leading modern data platform and the first database provider to IPO in over 20 years. The Search and AI Platform organization is extending Voyage's embedding models beyond first-party MongoDB Atlas.
Responsibilities
- Port and tune the model server that runs Voyage embedding and reranking models, improving inference performance, consistency, and runtime behavior across environments
- Productionize new Voyage models for delivery beyond first-party Atlas, owning packaging, configuration, and deployment workflows on AWS, Azure, GCP and more
- Design correctness, correlation, and performance validation that proves third-party deployments match first-party behavior
- Build operability into every surface: structured logging, metrics, diagnostics, and health checks with tools like Prometheus and OpenTelemetry
- Debug problems spanning model servers, containers, deployment configuration, and partner cloud environments
- Work alongside Voyage's model-serving teams and partner with GTM, SAs, TSEs, and strategic customers
Requirements
- 4+ years building backend, infrastructure systems, CI and build tooling
- Strong software engineering skills in languages such as Python or Go, with an emphasis on performance and reliability
- Experience with cloud environments (AWS, Azure, or GCP), containers, and Kubernetes
- Comfortable debugging across the stack - code, runtime, model servers, configuration, and external dependencies
- Familiar with concepts in ML model serving and inference runtimes
- Clear communication and success in ambitious, cross-team work
Nice to have
- Experience productionizing or deploying ML model servers or inference runtimes (e.g., vLLM, Triton, TGI)
- Experience with observability stacks like Prometheus, Grafana, or OpenTelemetry
- Experience shipping software through cloud marketplaces or partner / self-managed channels
- Contributions to open-source infrastructure for ML serving or deployment
Conditions
- Hybrid working model, based in Sydney