Principal Engineer

Valle Solutions; Colombia; Presencial; COP 120.000.000 - 210.000.000; Jornada completa

Descripción de la vacante

Valle Solutions is seeking a Principal Engineer to own outcomes across multiple systems and set technical direction for product surfaces and platform primitives.

You will work with campaigns, scoring, dealer-facing apps, feature flags, and CI, while standards for observability and production quality are raised across teams.

Formación

• Experience owning architecture across multiple systems and stacks.
• Proven ability to ship production systems with TS/Node, Java, or Rails.
• Experience applying AI tooling to improve shipping accuracy and reliability.
• Strong focus on observability, metrics, and release practices.

Responsabilidades

• Own architecture and delivery across more than one system, across TS/Node, Java, and Rails.
• Define and guide AI-enabled product features, scores, cohorts, or tool-calling.
• Document decisions with clear tradeoffs and teardown paths.
• Mentor Leads and Seniors; review architecture and PRs.

Conocimientos

TypeScript

Node.js

NestJS

Java

Spring

Rails

AWS

Datadog

AI tooling

Observability

Descripción del empleo

We're hiring a Principal Engineer, to own outcomes across systems. Not a single service. Not a Java-only seat.

You'll set technical direction on the product surfaces that move the business — campaigns, scoring, dealer-facing apps — plus platform primitives like feature flags and CI.

You'll sit with product and our scoring/agent layer (conversation, agents, MCP) when the work needs audiences, scores, or tool use. This is an IC role. You still write and review the hard code. You also write the architecture other teams build on.

What you actually do

• Own architecture and delivery across more than one system. Follow the problem, not the layer. Here that usually means TypeScript/Node (NestJS), Java (Spring), Rails, plus Datadog and AWS.
• Ship with AI as the default path — Cursor, agents, AI-green PRs. Teach the team what works. If humans keep leaving the same review comment, you update the agent policy. Design and review product features that use scores, cohorts, retrieval, or tool-calling when the product needs them. You are not hired to train models. You are hired to make those features reliable in production.
• Write the decision, not a slide. Company-wide calls (feature flags, reuse vs rebuild, new service vs extend) land as a short doc with a recommendation and a teardown path.
• Raise production quality: fail-closed defaults, observability next to release control, CI the rest of the org can copy.
• Mentor Leads and Seniors. Review architecture and PRs with veto when the design is wrong. Be the person other teams ping when the problem crosses a boundary.

What good looks like

• You've operated at principal, staff, or equivalent — owned a product or platform slice end-to end, not a ticket queue.
• You've shipped production systems on more than one stack. TypeScript/Node (NestJS) and/or Java (Spring) plus AWS is the usual path here. Rails fluency helps. You've used AI tooling to change how a team ships, not just your own speed. You can talk about prompts-as-spec, agent workflows, evals, and when not to use the model. You can design around an analytics or agent layer (scores, audiences, tool-calling) without needing to train models.
• You write clearly. Decisions get recorded. Tradeoffs are named.
• You mentor without being asked. You raise the floor.

Nice to have

• CRM, or campaign/orchestration domain
• Data-platform adjacency (warehouses, scoring APIs, PII-safe views)
• MCP, agent gateways, or IDE-agent workflows in a real org