Role Summary
The O2A team is building an autonomous multi agent generative AI system that interprets skan.ai process, mining data, generates SOPs. Reverse engineer business workflows and produces production ready code, tests and deploys with minimum human intervention.
Core Responsibilities:
LLM / Prompt Engineering
a. Design and Optimize:
- Multi-shot prompts
- Few shot and zero shot task prompts
- Chai of thought
- Program of thought
- Tool use prompts and agent action schemas
b. Create prompt templates for :
- SOP extraction
- Workflow reconstruction
- Code synthesis
- Test generation
- PR writing and review
Build multi-agent LLM workflows
a. Develop and maintain
- Planner agents
- SOP reasoning agents
- Code-Gen agents
- Testing agents
- CI/CD automation agents
- Implement memory state handling and agent-agent communication and task decomposition.
Integration and Orchestration:
- Use postman and similar tools for API debugging
- End point validation
- Workflow simulation
- Build integrations via APIs.
Code generation and automation testing:
- Generate production compliant Python/Java code modules and test suites.
- Validate code using sandbox runtime environments.
ML Ops/AI Ops
- Build and maintain embedding pipelines
- Automate telemetry and observability
Required Qualifications:
LLM and Prompt Engineering
- Expert Level prompt engineering
- Deep understanding of LLM reasoning. Context strategies and evaluation
AI Engineering
- Experience building LLM based applications using Python and Java
Hands-on experience with
- RAG
- Embeddings
- Agent frameworks (LangChain, AutoGen, etc.)
- SQL/MongoDB experience
- Excellent Python development skills
- Experience with Postman and APIs
- Experience with GitHub
- Familiarity with process mining (Skan.ai, Celonis etc..)
- Experience working in regulated environments (Banking/Finance/healthcare etc..)
- Familiarity with Audit processes aligned with regulatory requirements.
