Principal Platform Architect

TetraScience•San Mateo, CA

1d•$200,000 - $270,000

About The Position

TetraScience is building TetraOS, a Scientific OS for biopharma. The Tetra Platform provides the foundation for developing, delivering, and operating enterprise-grade, secure, compliant scientific data and AI capabilities. As the company scales, it is expanding industry partnerships and developer experience to support AI-native scientific data and workflows in production. The Principal Platform Architect will own the Tetra Platform’s architecture and evolution, as well as the partner integrations that extend it. This is an individual-contributor leadership role responsible for defining platform architecture aligned with product and business strategy and informing roadmap alignment decisions and trade-offs. The work spans enterprise data and agentic AI platforms, data and knowledge products, and developer products, including multi-tenant cloud architecture, durable workflows, AWS, Databricks, access controls, observability, search, semantic layers, knowledge and ontology products, and external developer platforms. The role’s scope includes enterprise tenancy, identity, compliance and control planes; scientific search; AI/ML operations; developer platforms and productivity; scientific use-case and workflow platforms; partner integrations; and cloud infrastructure. The architect will work across strategy, architecture, execution, and operations, bringing deep experience in several technology domains and balancing operational excellence with agility in a dynamic business context. In the first year, success includes streamlining authentication and authorization across platform layers; establishing an AI/ML infrastructure architecture and roadmap with operational telemetry, cost visibility, and governance; improving developer experiences, SDKs, templates, adoption, and reuse; defining and managing production and AI/ML service metrics and SLOs; implementing customer cost attribution and chargeback; building out data product and workflow primitives; evolving IDS toward open-standard schemas and strongly typed models with schema-on-write enforcement; and publishing partner reference architectures that enable at least one lab-instrument partner and one AI-model partner to onboard without bespoke engineering support.

Requirements

12+ years of software engineering experience, including at least 5 years at staff or principal level in a SaaS platform or data infrastructure context.
Demonstrated architectural and execution leadership in multi-tenant cloud architecture and data-intensive enterprise products at scale.
Earned expertise in at least two of the following: enterprise identity and authorization architecture, AI/ML serving infrastructure, search architecture, or data lake/warehouse/lakehouse architecture at scale.
Experience with enterprise identity and authorization architecture, SSO federation, or fine-grained authorization over data access, if counted among the required areas of expertise.
Experience building and operating model inference pipelines under production load, if counted among the required areas of expertise.
Experience designing and operating search platforms handling diverse query types across large structured or semi-structured datasets, if counted among the required areas of expertise.
Hands-on experience with data lake, warehouse, or lakehouse architectures at scale, including Delta Lake or Apache Iceberg, schema evolution, partition pruning, and query-performance versus storage-cost trade-offs, if counted among the required areas of expertise.
AWS infrastructure fluency with Kubernetes or ECS, including ability to investigate cost anomalies, identify root causes, and produce an action within the same week.
Ability to write, review, and defend architecture decisions, including RFCs, ADRs, technical trade-offs, and design reviews.
Strong cross-team communication and ability to write documents that produce alignment without requiring follow-up meetings for explanation.
Comfort working across strategy, architecture, and operations, including multi-year architecture direction and review of operational runbook gaps.
Experience across startups and large companies, with an understanding of trade-offs at different stages and scales.
Strong, diverse experience in operational excellence and agile execution in a dynamic business context.

Nice To Haves

Experience in regulated industries such as biopharma, medtech, or financial services, where compliance, cost-to-serve, and data residency are first-class architecture constraints.
Familiarity with scientific data platforms, ELN/LIMS systems, or laboratory informatics ecosystems, including constraints of instrument data.
Experience designing and operating internal developer platforms as a product, including roadmap, adoption metrics, and deprecation strategy.
Experience building partner integration programs at the architecture level, including connector SDKs, reference implementations, integration certification criteria, and self-service developer experience.
Exposure to laboratory instrument ecosystems, including proprietary data formats, on-premises agent deployment, and vendor certification workflows, or analogous hardware-adjacent integration work in medtech or industrial IoT.
Prior experience as a founding or early platform architect at a Series B–D SaaS company scaling to enterprise.

Responsibilities

Own the Tetra Platform architecture, its evolution, and the partner integrations that extend it.
Define platform architecture aligned with product and business strategy, and inform roadmap alignment decisions and trade-offs.
Design enterprise tenancy, identity, compliance, and control-plane capabilities, including SSO and fine-grained access controls for SQL data access.
Own scientific search architecture across keyword, semantic, and hybrid retrieval, including relevance standards, indexing pipelines, and reliable product infrastructure.
Set architecture for AI/ML operations, including model lifecycle, inference and training, Agentic IAM and developer experience, telemetry, observability, and production frameworks.
Build and evolve the external developer platform, scientific solution development experiences, integration tooling, SDKs, and adoption measures.
Improve developer productivity through toolchain ownership, local/production environment parity, and reducing friction from commit to deployment.
Accelerate adoption of the Scientific Use Case Library and develop data product and workflow platforms, IDS design standards, and the data access layer used by AI workloads and downstream pipelines.
Define integration architecture for laboratory instruments and AI model partners, including reference patterns, security boundaries, and self-service onboarding experiences.
Own production cloud architecture, cost governance, and observability from infrastructure signals through customer-visible service health.
Streamline authentication and authorization across data, compute, and agentic AI platform layers.
Establish AI/ML infrastructure architecture and a roadmap for inference and training, with operational telemetry, cost visibility, and governance.
Deliver streamlined developer platform experiences, SDKs, templates, adoption, and reusable contributions from Tetra, partner, and customer builders.
Roll out future-proof observability architecture and establish, monitor, and manage metrics and SLOs for production services and AI/ML workloads.
Operationalize customer chargeback attribution with finance and field teams.
Build out data products and workflow primitives with strong developer experience and operational scalability.
Evolve IDS toward open-standard schemas and encoding, strongly typed data models, and schema-on-write enforcement.
Publish reference architectures for lab-instrument and AI-model partner classes and enable partner onboarding without bespoke engineering support.
Write, review, and defend architecture decisions, including RFCs, ADRs, technical trade-offs, and design reviews.
Operate across strategic architecture direction and hands-on operational work, including production incidents and runbook gaps.