Descripción de la oferta
pOptiak is the modular operating system for enterprise AI. We extract the platform logic companies normally scatter across applications integrations, routing, security, memory, observability, prompt enrichment, and optimization into seven standardized layers that work with any model and any third-party tool, under one governed framework. /ppA platform like ours lives or dies on how well it connects to the customer’s world: their ERPs, CRMs, data lakes, document stores, internal APIs, and legacy systems. We’re hiring a founding architect to own that problem end to end. /ppbr/ppbThis /b bis /b ba /b bdeliberately /b bunusual /b brole. /b Most companies split it into three. You’ll do all three, and you should want to: /pullibBuild /b the platform primitives that make Optiak integration-ready out of the box. /lilibArchitect /b integration strategies with customers as a hands-on, technical solutions partner in pre-sales. /lilibDeploy /b those integrations inside real, regulated enterprise environments as a forward deployed engineer. /li /ulpIf you want a narrow brief and a clean spec, this isn’t it. If you want to own how an entire category of enterprise AI gets connected to the systems that run global companies, read on. /ppbr/ppbWHAT YOU'LL WORK ON /b /ppbBuild /b bthe /b bintegrations /b bplatform /b /ppYou’ll own Optiak’s bIntegrations /b bLayer /b and bPrompt /b bEnrichment /b bLayer /b the parts of the platform responsible for moving the right data to the right place, securely. /pulliDesign and ship pre-built connectors to major enterprise systems Salesforce, SAP, HubSpot, BigQuery, Google Drive, Slack, and the long tail of internal APIs so customers integrate in days, not quarters. /liliBuild the connection module to internal systems: ingestion, normalization, indexing, and permission-checked retrieval across both structured and unstructured data. /liliOwn the prompt-enrichment data flow the layer that decides what data a query needs, where it /li /ulp lives, and the most efficient way to retrieve it before a prompt reaches a model. /pulliDesign integration primitives for MCP and other emerging interoperability standards, so Optiak is /li /ulp the easiest platform to plug into any business ecosystem. /pulliMake the core architectural calls: when to copy data, when to federate, and when to query in place. /li /ulpbArchitect /b bwith /b bcustomers /b b(pre-sales) /b /pulliWork alongside the founders in pre-sales as the technical authority on integration: scope the customer’s data estate, surface constraints early, and design the right integration approach for each deployment. /liliTranslate messy enterprise reality legacy ETL, SSO/SAML, data residency, governance into a concrete architecture the customer’s own engineers trust. /liliBe hands-on, not slideware. Prototype, demo against real systems, and earn credibility with customer technical teams. /li /ulpbDeploy /b bin /b bthe /b bfield /b b(forward /b bdeployed) /b /pulliEmbed with design partners and early enterprise customers to take integrations from architecture to production inside their environment. /liliNavigate security reviews, compliance constraints, and the politics of getting production credentials the work that actually decides whether an AI deployment ships. /liliCodify what you learn into reusable connectors and patterns, and feed it straight back into the platform roadmap. /li /ulpbr/ppbWHAT WE ARE LOOKING FOR /b /ppbCore /b /pulli8+ years in backend, platform, or systems engineering, with deep experience in enterprise integrations and complex system architecture. /liliGenuinely hands-on. You write production bPython /b today and want to keep doing so this is not a role that drifts into pure advisory work. /liliStrong data-systems background: data lakes, APIs, batch and streaming pipelines, storage, and indexing. /liliProven track record designing systems used by large organizations at scale multi-tenant, secure by default, observable, and auditable. /li /ulpbJudgment /b /pulliYou think in systems, not services. You’ve designed platforms meant to be extended, not shipped /li /ulp once. /pulliYou make architectural tradeoffs across performance, security, and operational complexity under real constraints. /li /ulpbCustomer-facing /b /pulliYou’re comfortable in front of customers scoping, designing, and defending an architecture with their technical leaders and you’ve worked inside regulated environments (security, compliance, data residency). /liliYou can switch between writing code and explaining a design to a CISO in the same day. /li /ulpbAI-aware /b b(bonus, /b bbut /b bit /b bmatters /b bhere) /b /pulliFamiliarity with retrieval architectures, RAG, context assembly, and how data and integration choices shape LLM latency, cost, and quality. /liliExperience supporting AI products in production. /li /ulpbr/p
