Principal Engineer & Architect - AI Agent Plaform
Senior-most hands-on engineer and technical architect designing and building DevSavant's platform of managed AI agents that run business workflows for SaaS clients — owning integrations, data layers, APIs, infrastructure, security, and the agents themselves while coding the hard parts. Core stack: TypeScript/Node.js, React, PostgreSQL/Redis, LLM agents, RAG, and AWS/GCP.
- Compensation
- Not disclosed
- Location
- LATAM
- Work setup
- Not specified
- Seniority
- Principal
- Employment
- Not specified
- Experience
- 8+ years
- Category
- Architecture
- Posted
Skills
Must have
- 8+ years of production software experience, including recent hands-on ownership of a major system's architecture
- Deep experience in TypeScript and Node.js ecosystems
- Modern frontend experience (React, Remix / React Router v8, Tailwind, ShadCN)
- Relational ORMs (Drizzle) and validation tools (Zod)
- Databases (PostgreSQL, Redis) and queueing systems (PgBoss)
- Proven track record shipping and operating real production LLM agent systems with guardrails, evaluations, monitoring, and handling agent failure modes (loops, stalls, context drift, tool misuse)
- Deep understanding of RAG pipelines, chunking strategies, and vector data workflows
- Production infrastructure ownership: uptime, CI/CD, observability, and end-to-end incident handling
- Cloud architecture experience with AWS or GCP, including making and defending hosting/runtime decisions
- Production API integrations: webhooks, incremental polling, idempotency, retries, data normalization, and reconciliation
- Strong data modeling with relational and event-driven designs where data accuracy is the priority
- Multi-tenant SaaS security: tenant isolation, secrets hygiene, least-privilege access control, and audit logs
- Written-first leadership: clear, complete technical specifications enabling asynchronous remote execution
- Excellent written and spoken English and Spanish
Nice to have
- Official AWS or GCP Cloud Architect certifications
- Experience with agent frameworks beyond single-vendor stacks (e.g., OpenClaw, MCP, A2A)
- Experience with Infrastructure as Code and observability tools like Prometheus or Grafana
- Background in graph databases and knowledge graphs
- Early-stage startup experience building new product lines from 0 to 1
Full description
About DevSavant
DevSavant builds and operates managed AI agents / AI coworkers that run real business workflows in production for SaaS companies. We are turning that service into a scalable product: connecting client systems, normalizing their data, and running agent-driven workflows with human approval, dashboards, and full auditability.
Role Overview
You will be the senior-most engineer on the platform and its technical architect. You will design the integrations, data layers, APIs, infrastructure, security model, and the AI agents themselves. You present those designs to the Director of Engineering for sign-off, and then own building and shipping them alongside a distributed team of senior engineers.
How authority works here: you architect and propose, the Director of Engineering signs off on the direction, and you own execution, technical standards, and team direction from there. This is a hands-on architectural and technical leadership role. You will be writing code on the hard parts of the system, not managing HR functions or line-by-line supervision.
You will start by owning the architecture for the core platform (Managed AI Agents). As we expand into new business modules (Company Brain), you will take ownership of their technical architecture too, stepping in across projects to guide distributed engineering teams based on business priorities.
Responsibilities
- Whole-Platform Architecture: Design how client systems connect, how data is normalized and stored, how information is served to dashboards and AI agents, and how new client tools get integrated efficiently. You will be hands-on, building the primary framework skeletons yourself.
- Production AI Agent Systems: Handle agent design, tool usage, guardrails, evaluations, and monitoring for agents executing business workflows with human oversight.
- Cloud & Infrastructure Direction: Lead deployment architecture, CI/CD, observability, and cost management. The platform runs on containerized infrastructure today; deciding where the runtime should live long-term is one of your first major projects.
- Security & Multi-Tenant Architecture: Own tenant isolation, secrets management, least-privilege access, and audit trails for a platform handling financial data. You define the standards, and the team implements them.
- Technical Leadership: Set code standards, conduct reviews, and guide a distributed bench of senior engineers through clear written specifications.
- Client Engagement: Serve as the technical point of contact with clients and design partners when complex architectural discussions require it.