Senior Platform Engineer, ML Infrastructure (Xora Portfolio Company)

ABOUT ELEMYNT

Elemynt builds secure AI infrastructure for scientific and engineering R&D teams. Our platform helps organizations connect data, models, compute, and expert workflows in environments where reliability, traceability, and data control matter.

We are building a small, ambitious engineering team across Singapore and the United States to turn advanced scientific computing into production software that real technical teams can use.

ABOUT THE ROLE

This role owns the platform layer of Elemynt's product: the core services, including authentication and identity, authorization, multi-tenancy, and the middleware every request passes through.

You will build these services in Python, together with the delivery infrastructure that ships them and the telemetry that reports how deployments behave in the field. The work sits at the intersection of backend product development, release engineering, and observability, with roughly half of the time on the platform services themselves.

This is a deeply hands-on senior role. The platform is deployment-agnostic: it runs in the customer's cloud, on their own compute, or across a hybrid of the two, usually inside a secure environment we don't operate, on runtimes ranging from Kubernetes to HPC schedulers and bare metal. Identity has to integrate with each customer's existing provider, the software has to carry its own safeguards, and anything you will need to know about a live deployment has to be instrumented before the build ships.

WHAT YOU WILL DO

• Build authentication end-to-end: sessions, tokens, service identities, and integration with the customer's existing identity provider (OIDC, SAML, LDAP).
• Design the authorization and tenancy model and own it in code: roles, scopes, policy evaluation, and the tenant isolation that separates one customer's data and workloads from another's.
• Build the admin APIs for organizations, teams, and delegated access, and the middleware every request passes through: rate limits, quotas, and audit logging.
• Implement the safeguards that keep proprietary software and data protected in customer-managed installations: encryption at rest and in transit, secrets and certificate handling, and entitlement enforcement.
• Own the release path: CI/CD pipelines, quality gates, artifact registries, and versioned publishing, so a build that passes the gates is safe to ship.
• Package the platform for heterogeneous customer runtimes, with Helm charts, manifests, CLI installers, and infrastructure-as-code that make installs and upgrades repeatable.
• Operate ephemeral test environments that validate every change against a production-like installation before release.
• Collect usage metrics, logs, and traces from deployed instances, including networks with restricted egress, so deployment health is visible without access to the customer's environment.
• Instrument the platform's own services with the health signals and dashboards that make a deployment supportable and a failure quick to locate.

WHAT WE ARE LOOKING FOR

• Bachelor's or Master's degree in Computer Science or a related engineering field, and 6+ years building and shipping production software, with depth in the platform services a product is built on.
• Professional experience with Python as a primary language in production backend systems.
• Production ownership of authentication and authorization: session and token handling, issuance and validation, and a permission model (RBAC or similar) that other services depended on.
• Experience designing multi-tenant systems and the middleware around them: tenant isolation, permission modelling, rate limiting, quotas, and audit logging.
• Direct experience owning a release process other engineers depended on: CI/CD, quality gates, versioned artifacts, and deployment packaging with Kubernetes and Helm.
• A track record of instrumenting production systems and using the metrics, logs, and traces to debug incidents, with Prometheus, Grafana, OpenTelemetry, or equivalents.
• Experience shipping and supporting software in self-hosted, customer-managed, or on-premise installations.
• Experience at an early-stage company owning the trade-offs between scope, speed, and quality.

NICE TO HAVE

• License enforcement or entitlement systems, particularly the network license managers common in scientific and HPC software (FlexNet/FLEXlm, RLM).
• Telemetry or product-usage pipelines that report back from self-managed or customer-hosted installations.
• Experience across more than one runtime: cloud Kubernetes alongside HPC schedulers such as Slurm, or bare metal.
• Exposure to scientific computing, simulation, or other large-scale technical workloads.
• Go or Rust alongside Python.
• Time as an early engineer standing up a capability a company didn't previously have.

LOCATION

Singapore or United States. Work model is on site or hybrid, depending on location.

CLOSING NOTE

You do not need to tick every box. If this is clearly your kind of work, we would like to hear from you.