Job Specifications
Primary Skills : TypeScript/Node.js, Vertex AI, Gemini models, Gemini CLI.
Secondary Skills : GCP (Cloud Run, Functions, GKE, Pub/Sub, Cloud Storage, Firestore/BigQuery, Secret Manager),
Role Summary
We’re hiring an expert TypeScript/Node.js engineer to build production-grade GenAI services and developer tools. You’ll design and ship APIs, microservices, and prompt/LLM workflows on Google Cloud—leveraging Vertex AI, Gemini models, and the Gemini CLI—to deliver reliable, secure, and measurable business impact.
Design, implement, and own backend services in TypeScript/Node.js (Express/NestJS).
Build LLM-enabled features (retrieval, tool use, structured outputs) using Vertex AI and Gemini models; automate workflows with Gemini CLI.
Create, version, and evaluate prompts (zero/one/few-shot), including guardrails, safety filters, and offline/online prompt A/B tests.
Develop RAG and function/tool-calling pipelines; integrate vector stores and scalable embeddings on GCP.
Ship resilient APIs on Cloud Run/Functions/GKE, with Pub/Sub events, caching, and observability (Logging, Trace, Metrics).
Establish LLM evals (quality, latency, toxicity, hallucination) and telemetry/feedback loops for continuous improvement.
Harden services for security, privacy, and compliance (PII handling, data residency, secrets management).
Collaborate with product, ML, and UX to scope, estimate, and deliver features in iterative milestones.
Champion performance (p95 latency), cost efficiency (tokens/req), and reliability (SLOs, error budgets).
Mentor engineers on GenAI patterns, prompt engineering, and TypeScript best practices.
8+ years building backend systems with TypeScript/Node.js; expert in language/runtime internals, async patterns, and performance profiling.
Production experience with Google Cloud and Vertex AI (endpoints, model registry, evals) and Gemini models (text, multimodal, function calling).
Hands-on with Gemini CLI for rapid experimentation, prompting, and workflow automation.
Proven prompt engineering skill: prompt design/versioning, system prompts, safety strategies, and measurable eval frameworks.
Strong with REST/GraphQL, API contracts, schema validation (OpenAPI/Zod), and automated testing (Jest/Vitest).
CI/CD (Cloud Build/GitHub Actions), IaC (Terraform), and containerization (Docker).
Observability mindset: structured logging, distributed tracing, metrics/alerts.
Solid security fundamentals (OAuth2/JWT, least privilege IAM, secret rotation, threat modeling).
Tech Stack
TypeScript, Node.js
GCP (Cloud Run, Functions, GKE, Pub/Sub, Cloud Storage, Firestore/BigQuery, Secret Manager),
Vertex AI, Gemini models, Gemini CLI.