Make large language models reliable. You will design agentic workflows and prompt systems that produce safe, accurate, and cost-effective results, and you will prove it with evaluation, not vibes.

About the role

This is a frontier role. You will build the agents and prompt systems behind our AI products and client work, deciding when to use tools, memory, planning, and retrieval, and putting guardrails and evaluation around all of it. You should be equally comfortable crafting a prompt and reasoning about failure modes, cost, and latency.

What you will do

Design multi-step agent workflows with tool use, memory, planning, and retrieval.
Engineer, version, and systematically evaluate prompts against measurable criteria.
Build with frameworks such as LangChain, LlamaIndex, or native agent SDKs.
Stand up RAG pipelines and the evaluation harnesses that keep them honest.
Reduce hallucination and add guardrails for safety, privacy, and brand.
Tune for latency and token cost without sacrificing quality.
Instrument and observe agents in production and iterate on real traces.
Document patterns so the rest of the team can build reliable LLM features.

What we are looking for

Demonstrated experience building and shipping real LLM applications.
Strong prompt engineering instincts paired with a rigorous evaluation mindset.
Python and comfort integrating model APIs (OpenAI, Anthropic, DeepSeek, and similar).
A working understanding of RAG, tool use, function calling, and agent patterns.
The judgment to know when an agent is the right tool and when it is not.

Nice to have

Vector databases, MCP, structured output, and light fine-tuning.
Eval frameworks, tracing, and observability for LLM apps.
A point of view on agent reliability and where it breaks.

What success looks like

In 30 days: you have shipped an evaluated prompt or agent into a product.
In 90 days: you own an agentic feature with a real eval suite behind it.
In 6 months: our LLM features are measurably more reliable because of your systems.

What we offer

Remote-first culture with collaboration hubs across the USA, India, Singapore, and the UAE.
Work on applied AI and real products that ship to production, not slideware or endless pilots.
Senior mentorship, a dedicated learning budget, and sponsored cloud and AI certifications.
Competitive compensation, meaningful ownership, and a clear, fast path to grow.
A small, senior team where your work is visible and your decisions matter.

Apply for this role →See all roles