Selected Intern's Day-to-day Responsibilities Include
- Orchestrate AI agents: Design small agent teams (e.g., writer fact-checker editor) using frameworks like LangGraph or CrewAI. Define roles, prompts, and IO contracts, guardrails, and escalation paths.
- Build & refine RAG pipelines: Chunking strategies, indexing, retrieval quality, grounding with citations; balance recall/precision for Ayurveda knowledge.
- Evaluate & test: Create golden datasets and harnesses for unit/integration tests; measure quality, latency, cost; prevent regressions via automated checks.
- Rapidly iterate architectures: Diagnose failure modes (hallucination, drift, formatting errors) and adjust prompts, tools, or graph topology until outputs hit the bar.
- Instrument & observe: Add tracing/telemetry for runs, token/latency budgets, and quality dashboards.
- Collaborate for impact: Translate growth goals into agent flows; ship internal tools that enable content production, SEO experiments, and campaign support.
- (Nice) Wire up tools via MCP: Use the Model Context Protocol (MCP) to connect agents to internal data/services.
About Company: Consuma's Rapid Research platform is a real-time insights and research engine built for brands and businesses to exercise while making key decisions. Traditionally, consumer research has been long, expensive, and inconsistent. Consuma aims to revolutionize this industry - we've been working with hundreds of brands, delivering insights and research within minutes, at a fraction of the cost.
