Principal Engineer & Architect - AI Agent Plaform

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.

Requirements

• 8+ years of production software experience, including recent hands-on ownership

• of a major system's architecture.

• Modern Tech Stack Expertise: Deep experience in TypeScript and Node.js

• ecosystems, modern frontends (React, Remix / React Router v8, Tailwind, ShadCN),

• relational ORMs (Drizzle), validation tools (Zod), databases (PostgreSQL, Redis),

• and queueing systems (PgBoss).

• Production LLM Agent Systems: Proven track record shipping and operating real

• LLM agent systems, not just prototypes. You have built guardrails, evaluated agents

• prior to launch, monitored them in production, and handled agent-specific failure

• modes like loops, silent stalls, context drift, and tool misuse.

• RAG & Engineering Depth: Deep understanding of RAG pipelines, chunking

• strategies, and vector data workflows.

• Production Infrastructure Ownership: Experience managing uptime, running

• CI/CD, setting up observability, and handling production incidents end to end.

• Cloud Architecture (AWS or GCP): Proven experience making and defending

• decisions around platform hosting and runtimes.

• Production API Integrations: Experience with complex third-party API

• environments involving webhooks, incremental polling, idempotency, retries, data

• normalization, and data reconciliation.

• Data Modeling: Strong experience with relational and event-driven data designs

• where data accuracy is the priority.

• Multi-Tenant SaaS Security: Practical experience with tenant isolation, secrets

• hygiene, least-privilege access control, and audit logs.

• Written-First Leadership: Ability to write clear, complete technical specifications that

• allow remote engineering teams to execute high-quality work asynchronously.

• Communication: Excellent written and spoken English and Spanish.

Nice-to-haves

• Official AWS or GCP Cloud Architect certifications.

• Experience with agent frameworks beyond single-vendor stacks (such as OpenClaw,

• MCP, or 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→1.