About the role

Design, build and run AI agents that help the team brief demand, draft and check configuration, explain decisions, and review and test code. Make them safe, measured and worth their cost. Work confidently with client stakeholders and say when AI is the wrong answer.

Key responsibilities

- Build AI agents with tool use, retrieval, memory and orchestration
- Connect agents to company tools and data through MCP and APIs, with least-privilege access
- Build evaluation sets and automated evals (quality, safety, latency, cost)
- Put guardrails in place (personal data screening, permissions, audit trails, human approval)
- Monitor agents for quality, drift, failures and cost
- Choose models and hosting (Claude, AWS Bedrock, Google Vertex AI) based on evidence
- Write production code with tests and CI/CD
- Write ADRs for key AI decisions
- Coach the team on good AI practice

Mandatory skills

1.



AI features or agents used by real users in production
2. Agent patterns: tool calling, retrieval, memory, multi-step workflows
3. Building evaluation sets and automated checks
4. Guardrails: prompt-injection defence, personal data handling, human-in-the-loop
5. Robust coding in Python, TypeScript or Kotlin, with tests and CI/CD
6. Integrating AI with enterprise tools through APIs (MCP preferred)
7. Experience with Anthropic, OpenAI, AWS Bedrock or Google Vertex AI
8. Reducing cost and latency of AI features
9. Explaining AI clearly to non-technical stakeholders
10. Ownership mindset: owning a problem from requirement to production outcome

Good to have MCP servers, Claude Code, Confluence retrieval, red teaming, AWS/GCP, Kafka, rules-heavy platforms

📌 AI Technical consultant (Bengaluru)
🏢 Moolya Software Testing
📍 Bengaluru