Corti builds inference-heavy healthcare AI, evident from a production-focused stack (PyTorch, TensorFlow, NVIDIA Triton, vLLM, Kafka) paired with infrastructure challenges around low-latency serving and 100k concurrent stream scaling. The org is engineering-led with active adoption of GitOps and Model Context Protocol, signaling a shift toward operational maturity and AI agent tooling. Pain points cluster around production reliability (model drift, latency) and multi-tenant infrastructure — common for APIs moving from pilot to enterprise workloads.
Tech Stack 35 technologies
Core StackGo Kubernetes Docker AWS Python MLflow Kubeflow ArgoCD FastAPI PyTorch Apache Kafka Java GraphQL CockroachDB Elasticsearch Azure DevOps Jenkins Kafka gRPC Drata TensorFlow Loki Grafana Azure GCP DVC vLLM NVIDIA Triton Inference Server WebSockets WebRTC+2 more
AdoptingGitOps Replit Model Context Protocol
What Corti Is Building
◆Challenges
- Clinicians overwhelmed
- Medical knowledge outpacing human capacity
- Lack of access to quality healthcare
- Scaling infrastructure for high concurrency
- Maintaining multi-tenant cluster reliability
- Model reliability in production
- Model performance issues
- Partner ecosystem growth
- Scaling partner integrations
- Low-latency inference
▲Active Projects
- Automation tools for developer experience
- Speech recognition and text generation backend
- Low-latency ai integration pipeline
- Infrastructure scaling for 100k concurrent streams
- Ci/cd pipeline development
- Ml model deployment automation
- Model governance implementation
- Agent store growth
- Multi-tenant kubernetes setup
- Certified solutions partner network
Hiring Activity
15 roles · 9 in 30d
Department
Seniority
Company intelligence
