InfrastructureFull-Time · Remote-Hybrid
Cloud and MLOps Architect
Hypotenuse Analytics
CompanyHypotenuse Analytics
LocationRemote-Hybrid
TypeFull-Time
Experience1-3 YOE
About Hypotenuse Analytics
Hypotenuse Analytics is building an AI-powered monitoring platform across three domains: structural health monitoring of physical infrastructure using IoT sensors and satellite data, real-time video analytics for surveillance and anomaly detection, and synthetic media detection for KYC, enterprise communications, and media verification. The three products share a common data and inference layer, which is what makes the architecture interesting and the engineering genuinely hard.
We are at seed stage, running our first pilots in 2026. The team is small. Everyone here touches real problems from day one.
A Note on AI
AI tools are part of how we work. Use them. What we evaluate is judgment and craft, and our bar is higher because of it. Low-effort generations with no thinking behind them are easy to spot and will cost you the role.
About the Role
The three products we are building all converge on the same core problem: getting model inference to happen fast, reliably, and cheaply at scale. Video streams need low-latency GPU inference. SHM needs continuous ingestion from IoT sensors and satellite data APIs. Deepfake detection needs a high-throughput endpoint that financial and enterprise clients can call from their own systems.
You will be the first dedicated infrastructure hire. You will design the AWS architecture for all three pipelines, advise on GPU instance selection, build the MLOps layer so the ML team can ship and monitor models without depending on you for every deploy, and keep costs under control as we move from pilots to production.
What You Will Work On
- Design the AWS cloud architecture from scratch: VPC layout, IAM structure, compute, storage, and networking across our three product pipelines.
- Own GPU instance selection for model inference workloads. Video analytics, deepfake detection, and SHM each have different latency and cost profiles. We want someone who can advise on the trade-offs across instance families.
- Build and maintain the MLOps layer: model registry, experiment tracking, versioned training pipelines.
- Set up inference serving for our CV models: autoscaling endpoints, staged deploys, and routing between model versions.
- Build ingestion pipelines for IoT sensor streams and satellite data APIs.
- Manage our early-phase AWS Rekognition setup and plan the migration to self-hosted models as volume grows.
- Set up observability: model performance dashboards, infrastructure metrics, and alerting.
- Build CI/CD for ML with validation gates, staged rollouts, and rollback procedures.
Core Stack
AWS (primary)TerraformKubernetes on EKSMLflowDVCDockerGitHub ActionsSageMakerKafka or SQSPython
What We Are Looking For
- Solid hands-on AWS experience in a production environment.
- Some experience with GPU workloads on AWS, even if not at scale.
- Familiarity with MLOps tooling and the idea of building a platform the ML team can actually use without hand-holding.
- Infrastructure-as-code mindset. Terraform or equivalent.
- Comfortable owning an entire cloud stack with limited senior oversight.
- 1-3 YOE
Good to Have
- Experience with IoT data ingestion, video stream processing, or GPU inference optimisation.
- An infra project on GitHub that is actually deployed somewhere.
Ready to Shape the Future?
Submit your credentials and resume through our online recruitment portal.
Careers Team · [email protected]