P
AI Engineer Remote (United States) · View role
Point11
5 hours ago
Full-time
On-site
Miami, Florida, United States
We are a team building the infrastructure for the agent era. A decade ago, the brands that leaned into mobile won. The same shift is happening with agents: how people find businesses, ask questions, and buy is moving from static pages into conversations. 800 million people already discover brands through agent search, and who gets chosen is being decided right now.
Build the agents and model pipelines at the heart of Point11: the digital twins that stand in for a company's customers, the campaigns that question them, and the Chat and Voice agents that answer real visitors. You own prompts, evals, retrieval, and tool use from the first experiment to production. Every interaction has to feel intelligent, and it has to happen in .11 seconds. Small team, high bar. Founded by ex-Googlers. We hire accordingly.
What Your First Year Looks Like
Build the eval suite that tells us, every day, whether a customer's digital twins still answer the way its real customers do. Cut the latency of a Voice agent's turn while raising its answer quality, measured on real calls. Design the agent that turns a campaign's findings into the next actions a brand should take, and prove it with outcomes. What the Day-to-Day Feels Like
Ship LLM features end to end: prompts, retrieval, tool use, structured outputs, evals Design evaluations that catch regressions before customers do Choose the right model for each job and know why, on cost, latency, and quality Instrument every agent so we can see what it did and why Work with product and Forward Deployed Engineers to turn what customers need into working agents About You
You've shipped LLM-powered features to real users, not just demos You're strong in TypeScript or Python and comfortable across the stack You've built evals and you trust them more than vibes You think in systems: retries, timeouts, costs, and failure modes You live in the United States. This role is remote, with regular time in Miami. Even Better If You Have
Voice or real-time conversational agents Retrieval at scale: embeddings, vector search, hybrid ranking Agent frameworks, MCP, or tool-calling protocols Research experience in NLP or ML What We Offer
Your agents are the product customers see Remote in the United States, with time together in Miami The newest models, used in production the week they land A direct line to the founding team on product An annual learning budget for conferences and development
#J-18808-Ljbffr
Build the eval suite that tells us, every day, whether a customer's digital twins still answer the way its real customers do. Cut the latency of a Voice agent's turn while raising its answer quality, measured on real calls. Design the agent that turns a campaign's findings into the next actions a brand should take, and prove it with outcomes. What the Day-to-Day Feels Like
Ship LLM features end to end: prompts, retrieval, tool use, structured outputs, evals Design evaluations that catch regressions before customers do Choose the right model for each job and know why, on cost, latency, and quality Instrument every agent so we can see what it did and why Work with product and Forward Deployed Engineers to turn what customers need into working agents About You
You've shipped LLM-powered features to real users, not just demos You're strong in TypeScript or Python and comfortable across the stack You've built evals and you trust them more than vibes You think in systems: retries, timeouts, costs, and failure modes You live in the United States. This role is remote, with regular time in Miami. Even Better If You Have
Voice or real-time conversational agents Retrieval at scale: embeddings, vector search, hybrid ranking Agent frameworks, MCP, or tool-calling protocols Research experience in NLP or ML What We Offer
Your agents are the product customers see Remote in the United States, with time together in Miami The newest models, used in production the week they land A direct line to the founding team on product An annual learning budget for conferences and development
#J-18808-Ljbffr