
Node TS + React TS + LLM Agents | Full Stack Developer at S.A.G.A Corp. You're built for big things, not endless marketing tasks. At SAAGA Solve, our intelligent agents manage all the tricky parts of getting noticed on search and social media, so established businesses and growing startups can truly focus on what they do best - innovating, building, and serving their customers.. We’re hiring a senior full-stack engineer to help us build AI-powered products end-to-end. You’ll own features from rough idea to production: architecture, implementation, testing, release, and iteration. The work spans modern web (Node.js, React, Next.js, TypeScript) and LLM agents (tooling, function calling, integrations, evaluation, and reliability).. You’ll thrive here if you: (a) write clean, production-grade code fast, (b) treat tests and monitoring as part of “done,” (c) think like a founder—challenge assumptions, propose better paths, make decisions with incomplete data, and ship. We’re an A-player team; we expect proactive ownership, direct communication, and the courage to push back when something isn’t right.. What success looks like:. In first weeks days: ship a user-visible feature, add/extend an agent tool, and land solid PRs with strong tests. . In first Q days: lead a small project across front-end, back-end, and an LLM integration (OpenAI/Anthropic/Gemini or Vercel AI SDK/OpenRouter), with metrics and basic evals. . Be a go-to owner for a product slice, raising code quality, performance, and reliability while mentoring others on agent patterns.. 12+ years. professional software engineering experience delivering production systems. . Full-stack depth:. Node.js, React.js, Next.js, and . TypeScript. at an expert level (types as design, not decoration). . AI/LLM agents:. hands-on experience building agents with function/tool calling, multi-tool orchestration, retries, cost/latency trade-offs, and safety/guardrails. Comfortable with prompt design, schema design, and deterministic interfaces. . APIs & SDKs:. strong with OpenAI, Anthropic, Google Gemini; plus nice to have one of . Vercel AI SDK. or . OpenRouter. . Able to design clean abstraction layers over external models/tools. . Systems & product thinking:. convert fuzzy requirements into simple, resilient architectures. Bias to ship; know when to spike, when to harden. . Testing discipline:. writes your own . unit, integration, and happy-path/end-to-end tests. ; partners with QA for edge-cases and complex scenarios. “It works on my machine” is not done. . Quality in production:. observability (logs/metrics/traces), feature flags, canary/rollbacks, performance profiling, and error budgets are familiar tools, not buzzwords. . Data & storage:. practical experience with REST/GraphQL, Webhooks, background jobs/queues, caching, and at least one SQL database. Comfortable with pagination, idempotency, and rate-limit strategies. . Security & privacy:. sound judgment around secrets, authZ/authN, PII handling, and least-privilege by default. . AI-assisted development:. enthusiastic and effective using AI for coding/review/testing, while maintaining rigorous standards and architectural sense. . Collaboration:. crisp written communication, clear PRs, structured RFCs, and constructive code reviews. . Founder mindset:. proactive, decisive, comfortable with uncertainty. Preferred: familiarity with . Y Combinator. content/principles (default alive, talk to users, build/measure/learn). . Nice to have (not required):. RAG pipelines, vector DBs, lightweight eval harnesses, prompt/version management, streaming UX, and cost telemetry.. Company Location: United Kingdom.