SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Platzi is the largest technology school in the Hispanic world, serving over 7 million students and 4,000 companies. The company trains individuals and organizations in AI, Cybersecurity, English, Programming, and Leadership, with a mission to make Latin America a technology superpower.
As a Senior Software Engineer on the Learn platform, you will be an engineer first—building and owning production systems powered by AI. You report to the AI Lead and work on orchestration, evaluations, and model decisions that directly impact how millions of students learn in the product. Seniority here means sound judgment and the ability to defend architectural choices, not distance from code. You care about system costs and maintainability over the long term, not just shipping features.
Key Responsibilities:
- Build and own AI features shipped to students, managing performance, reliability, cost, and failure recovery in production.
- Design orchestration systems: agents, tools, and workflows built to be debugged and understood, not just demoed.
- Integrate evaluations into your workflow to track improvements, regressions, and trade-offs.
- Select the right technical approach for each problem—LLMs, classic ML models, or deterministic code.
- Make defensible architecture decisions and reject complexity that doesn't justify itself.
- Raise engineering standards through code review, technical standards, and early intervention on problematic decisions.
Requirements:
- Minimum 5 years building software in production, including owning a system through real scale or a migration.
- Strong fundamentals in data modeling, APIs, performance, concurrency, and failure modes.
- Proven experience taking LLM-powered systems to production (not prototypes); understanding where they fail.
- Comfortable working across frameworks like Pydantic, Agents SDK, and A2A; able to justify framework choices.
- Experience with classic machine learning and judgment about when models outperform prompts.
- A portfolio piece—a product you built that can be reviewed and used.
- Fluent in both English and Spanish.
