AI Engineer

Alyssum Global Services is looking for an AI Engineer to build, deploy, and operate production AI systems that address real business needs. This is a production ownership role, not a research-only position. You will take problems from business requirements through design, launch, and ongoing operation. The role is based in Noida, Sector 90, and follows a hybrid work mode.

WHAT YOU WILL DO

• Build agentic workflows and tool-using agents that reliably complete multi-step tasks, with retries, error handling, and human approval steps, integrated over REST APIs, webhooks, and MCP servers.
• Own the RAG stack end to end: ingestion, chunking, embeddings, retrieval, re-ranking, and generation.
• Measure and improve output quality: own the evaluation framework and manage prompts as versioned, tested artifacts.
• Run AI systems in production (about a third of the role), including monitoring, inference costs, and model upgrades.
• Secure AI systems through prompt-injection defenses, sensitive-data safeguards, and agent-permission controls.
• Work with stakeholders to identify where AI can reduce manual work, scope and estimate solutions before building, support adoption, and document what you build.

WHAT YOU BRING

• 3+ years of software engineering experience, including at least 1 year building AI/LLM systems that reached production (mandatory).
• Strong Python skills, including async code for concurrent  API calls.
• Hands-on experience with a leading LLM API (Anthropic Claude, OpenAI, or Google Gemini), including function calling, structured output, and context management.
• Experience building a production RAG system end to end, plus practical experience evaluating LLM output.
• Integration experience with REST APIs, webhooks, and OAuth, plus solid fundamentals in Git, SQL, Docker, CI/CD, and one cloud platform (AWS, Azure, or GCP).
• Bachelor's degree in Computer Science or a related field, or equivalent hands-on experience.
• Strong ownership, business sense, healthy skepticism of model output, and fluent English.

Additional advantages: MCP server development; agent frameworks (LangGraph, Claude Agent SDK, LlamaIndex); vector databases; low-code automation (Monday.com, n8n, Make, Zapier); TypeScript/Next.js; open-weight models; fintech or other regulated environments.