
Job Description
Ovii's Interpretation of the Role
Metaforms seeks a senior AI engineer to design, build, and continuously improve production‑grade AI agents that power large‑scale market‑research surveys. The role blends applied LLM work with systems engineering, demanding high ownership of agent harnesses, evaluation pipelines, and reliability tooling.
Role Snapshot
- Senior AI Engineer
- On‑site in Bengaluru
- Full‑time
- High‑ownership AI systems
- Production LLM/agent pipelines
- Python‑centric development
Must-Have Requirements
- Python
- Frontier model APIs (Anthropic, OpenAI, Gemini)
- Context engineering
- Debugging complex non‑deterministic failures
- production AI agent systems
- LLM/agent pipelines
Nice-to-Have Signals
- LLM observability and eval tooling (Braintrust, Langfuse, LangSmith, Weave, promptfoo)
- Semantic parsing / DSLs
- Computer‑use or browser agents
- Human‑in‑the‑loop workflows
- Go
- TypeScript
- LLM observability tooling
- semantic parsing and DSLs
- computer‑use agents
Work Setup
- Location: Bengaluru, India
- Work mode: ONSITE
- Employment type: Full-Time
Not Specified in JD
- Visa sponsorship
- Salary range
- Remote eligibility
- Coding test
- Portfolio
What You'll Likely Work On
- Own and evolve the agent harness that orchestrates planning, tool use, and self‑repair
- Research and implement long‑context handling and cascading‑error mitigation in multi‑step pipelines
- Define structured evaluation rubrics and build continuous monitoring, tracing, and failure‑mode analysis for agents in production
- Create tooling for domain experts to update skill files, eval sets, and knowledge bases
- Develop regression suites, golden datasets, and LLM‑as‑judge pipelines to catch regressions before deployment
- Design human‑in‑the‑loop review workflows for high‑stakes survey outputs
Good Fit If You Have
- Enjoys fast‑paced environments with zero‑bureaucracy
- Thrives when decisions have direct architectural impact
- Comfortable collaborating with domain experts and leadership
Skills
- Python
- Frontier LLM APIs (Anthropic, OpenAI, Gemini)
- Context engineering
- Debugging non‑deterministic systems
- Go or TypeScript (plus)
- LLM observability & eval tooling
- Semantic parsing / DSLs
- Computer‑use / browser agents
- Human‑in‑the‑loop workflows