Software Engineer, AI/ML New York City

Kyahq; New York (NY); Hybrid; USD 180,000 - 240,000; Full time

Benefits offered by this job

Health & dental insurance

Lunch & dinner

In-person energy

Competitive cash and equity

Job summary

Kya is hiring a founding software engineer to build the AI agent platform that powers customer success across real accounts. You’ll own systems end to end—from the agent runtime to memory, retrieval, and evaluation harnesses—and ship production services used by customers and revenue.

You’ll work directly with the founders and shape the architecture and roadmap as the product matures. This is a hands-on role in a fast-growing startup with a strong emphasis on safe, scalable AI systems.

Qualifications

• 4+ years shipping production software with real users.
• Fluency with the modern AI stack including LLM APIs.
• Strong fundamentals in distributed systems and API design.
• Bias toward shipping, iteration, reliability and safety.

Responsibilities

• Design and build the agent platform with planning, tool use, orchestration and guardrails.
• Build memory and retrieval layer so agents learn playbooks per customer.
• Integrate and orchestrate LLMs, own prompting, context engineering, routing, costs and latency.
• Stand up evaluation and observability systems to detect improvements and regressions.
• Ship production services: APIs, pipelines, and infrastructure behind agents.
• Use Kya to build Kya and evolve the platform.

Skills

Production software

LLM stack

API design

Distributed systems

Bias toward shipping

Agent frameworks

Tools

LLM APIs

RAG

Vector databases

Evaluation harnesses

Job description

About Kya

Kya is your customer success agent. We build AI agents that onboard, activate, and expand customer accounts while teams focus on the relationships, turning customer success into an operating system for growth. We're an agent-native company with a high bar for everything, and we believe a job was never meant to be the busywork. It was meant to be the thinking and the building. Every action our product takes makes it better at the next one, and we're hiring the engineers who make that compounding loop real.

The role

You'll build the AI that runs Kya. That means the agent runtime, the memory and retrieval systems, the evaluation harnesses, and the model orchestration that lets agents take real actions on real customer accounts safely and reliably. This is a founding engineering role: you'll own systems end to end, make architectural calls that stick for years, and ship them to production where customers and their revenue depend on them. You'll work directly with the founders and the people closest to our customers, and the loop from idea to shipped agent is short by design.

What you'll do

• Design and build the agent platform: planning, tool use, orchestration, and the guardrails that keep agents acting safely on customer data.
• Build the memory and retrieval layer that lets agents learn a customer's playbooks and get smarter with every account.
• Integrate and orchestrate LLMs (Claude, OpenAI, and others), and own prompting, context engineering, routing, and cost and latency tradeoffs in production.
• Stand up the evaluation and observability systems that tell us whether an agent is actually getting better, and catch regressions before customers do.
• Ship production services: the APIs, pipelines, and infrastructure behind agents that take millions of actions across our customers' stacks.
• Use Kya to build Kya, and turn what you learn into the next version of the platform.

What we’re looking for

• A strong software engineer who ships. You write production code, own systems end to end, and care about correctness, latency, and cost in the real world.
• 4+ years building and shipping production software, with hands‑on experience putting LLM‑or ML‑powered features in front of real users.
• Real fluency with the modern AI stack: LLM APIs, RAG and retrieval, agent and tool‑use patterns, and evaluation.
• Strong fundamentals in distributed systems, data pipelines, and API design.
• A bias toward shipping and iterating fast, and the judgment to know where reliability and safety cannot be cut.
• Genuine excitement about agents, and about being early at a company building them.

Nice to have

• Experience as an early or founding engineer at a startup.
• Experience with agent frameworks, vector databases, or building evaluation harnesses from scratch.
• Familiarity with CRM, support, or customer success data and integrations (Salesforce, HubSpot, Zendesk, Slack).
• Open-source work, side projects, or shipped agents you can show us.

Benefits

• Health & dental insurance
• Lunch & dinner
• In-person energy
• Competitive cash and equity compensation