AI Engineer, Agents & Intelligence

Scalio is building an AI marketing team for millions of small businesses. We’re looking for people who take ownership, build thoughtfully, and turn their work into measurable customer outcomes.

AI Engineer, Agents & Intelligence

Scalio | Bengaluru, India | On-site | Full-time

Compensation: ₹30–60 LPA annual fixed salary plus equity. Equity terms discussed during the hiring process.

We're looking for an AI Engineer to build the intelligence layer behind it: agents that understand a business, remember context, use tools, make decisions and execute marketing workflows.

This is not a prompt-engineering role and not an ML research role.

You'll work on:

• AI agents and multi-step workflows
• Long-term memory for businesses and users
• RAG, retrieval, embeddings and vector search
• Tool/function calling and API orchestration
• Context engineering and model routing
• AI evaluation, tracing and reliability
• Production LLM systems optimized for quality, latency and cost

What we're looking for

• Strong software engineering fundamentals
• Excellent Python
• Approximately 2–5 years of engineering experience, though exceptional candidates can be outside this range
• Hands-on experience building and shipping LLM/AI systems
• Experience with RAG, embeddings, vector databases or search
• Experience building agents, tool-calling systems or agentic workflows
• Understanding of memory, context management and stateful AI systems
• Strong backend fundamentals: APIs, databases, async systems and production infrastructure
• Ability to take an ambiguous problem from idea → architecture → production
• Strong curiosity and ability to quickly adopt new AI models and techniques

Strong plus

• LangGraph / LlamaIndex / OpenAI Agents SDK / PydanticAI or similar
• MCP
• Hybrid search / reranking
• LLM evaluation and observability
• Multimodal AI
• Redis / PostgreSQL / pgvector
• Docker / cloud infrastructure
• Open-source contributions or impressive personal AI projects
• Heavy use of Claude Code, Codex, Cursor or similar AI coding tools

What we care about less

• Your degree
• LeetCode performance
• Having worked at a famous company
• Knowing every AI framework
• Exact number of years of experience

We care much more about what you've actually built and shipped.

If you've built something technically impressive, tell us about it.

Required written questions

1. What is the most technically complex AI/LLM system you have personally built and shipped to production? Briefly explain what you built, your specific contribution, scale/users, and one difficult problem you solved.

2. Share a link to something you have built: GitHub, demo, product, technical article, paper, open-source project, etc. What part did you personally build?

3. Suppose an AI agent needs to remember a business's important information across 6 months of conversations and actions. How would you design its memory system? Briefly describe what you would store, where you would store it, and how you would retrieve/update it.

4. Describe an AI agent or agentic workflow you've built. What tools could it use, how did you manage state/context, and what happened when the model made a wrong decision?

5. How have you evaluated whether an LLM/RAG/agent system actually improved? Give one example of a metric, test set, or evaluation process you have used.

6. You inherit a production RAG agent with 90% retrieval recall but poor answer quality. What would you investigate first?

Application — all six answers required

Apply through our Google application form: https://docs.google.com/forms/d/e/1FAIpQLSfwGInRwRVxoMUzsFKTjLSOdJbFyeV2y21liQgALugUMIBhUw/viewform

Provide your name, email, phone number including country code, résumé/CV link, relevant experience, availability to work full-time on-site in Bengaluru, and answers to all six questions above. Include relevant links and clearly explain your personal contribution. Use concrete examples; generic answers will not stand out. Please omit confidential employer or customer information.

We care more about demonstrated work and results than degrees, company names, or familiarity with every tool.