Senior AI Product Engineer (Agent Systems) - Nas.com

Nas Company; Singapore; Hybrid; SGD 120,000 - 170,000; Full time

Benefits offered by this job

Stock options

Hybrid work in Singapore

Relocation support

Mental-health funds

Company lunches

Grab allowance

Yearly retreats

Job summary

Nas.com is seeking a senior engineer to own the agent systems behind Nas.com. You will design and run the runtime that keeps agents accurate, fast, and safe, while owning backend integrations that bring customer data into the platform.

You will shape how our agents are built and how they interact with data across the stack. You have shipped LLM agents for real users, built orchestration cores, and can demonstrate improvements in accuracy, latency, and cost.

Qualifications

• 5+ years of software engineering with backend and systems design.
• Shipped and operated LLM agents for real users.
• Built or extended the orchestration core: state, tool execution, retries, recovery.
• Built agent evals to improve accuracy or task success.
• Cut agent latency or cost in production and demonstrated impact.
• Able to structure ambiguous problems into working outcomes.

Responsibilities

• Build the core of our agents, including loops, tool execution, context and memory.
• Design safe, permissioned guardrails and approval flows for real-money actions.
• Measure quality with evals from real workflows and failures.
• Reduce cost and latency while maintaining quality through architecture choices.
• Set technical direction for agent structure, evals, and observability.
• Own delivery end-to-end from design to production in customer environments.

Skills

Backend

System design

LLM agents

Agent evals

Latency reduction

Ambiguity to solution

Python

TypeScript

AWS

LangChain

LangSmith

MongoDB

ClickHouse

Tools

LangChain

LangSmith

MongoDB

ClickHouse

AWS

Python

TypeScript

Job description

We are hiring a senior engineer to own the agent systems behind Nas.com. Our agents watch ad accounts, social channels, and the market around the clock, build a live memory of each business, and act on what they find once a human says yes. Your job is to make them reliable, fast, cheap to run, and safe to trust.

About Nas.com

Nas.com builds AI that does the work of growing a business, not just answers questions about it.

• Agents (nas.com/agents) is our AI agent platform for marketing teams. It connects ads, social, CRM, content, analytics, and transaction data into one marketing brain for each company. Agents then work on it 24/7: watching ad accounts and social channels, improving ad performance, monitoring reputation and competitors, and flagging what needs a human decision.
• Solopreneur (nas.com/solopreneur) is our platform for solo entrepreneurs. With help from AI agents, they can set up a store, generate marketing materials, connect to ad platforms for traffic, and collect payments worldwide, all in minutes.

We have raised US$40 million, most recently in a round led by Khosla Ventures and 500 Global. We are based in Singapore and build for customers worldwide.

The Role

This is a hands-on role, and you are responsible for delivery. Our agents run in production for paying businesses, with real money on the line, and the results are yours to own. You will decide how our agents are designed and built, own the runtime that keeps them accurate, fast, and efficient, and use your own judgment to make the AI smarter with every release. You will also own the backend integrations that bring customer data into the platform. Agent systems come first, but you will reach across the stack when the problem needs it.

What You'll Do

• Build the core of our agents. Agent loops, tool execution, context and memory, and workflows that survive failures. Add multi-agent coordination only where it helps.
• Make agents safe to act. Design the permissions, guardrails, and approval flows for agents that spend real money and write to real systems.
• Make quality measurable. Build evals from real workflows and real failures. Measure accuracy, tool use, and end-to-end task success, and catch regressions before release.
• Cut cost and latency without losing quality. Test model routing, retrieval, context, caching, and concurrency against real workloads. Track cost per successful task and tail latency.
• Set the technical bar. Define how we structure agents, evals, and observability. Lead design reviews and help the team make better calls.
• Own delivery end to end. From design to a solution that is shipped and running inside a customer's business. When the same pain shows up twice, fix the system, not the symptom.

Who You Are

• 5+ years of software engineering, with strong backend and system-design fundamentals.
• You have shipped and operated LLM agents for real users, and you own the decisions and failures that came with them.
• You know what agent frameworks do under the hood. You have built or extended the orchestration core yourself: state, tool execution, retries, recovery.
• You have built agent evals and used them to improve accuracy or task success.
• You have cut agent latency or cost in production and can show how you measured it.
• You take an ambiguous problem, give it structure, and drive it to a working outcome.

Bonus Points

• Time at both a large tech company and an early-stage startup.
• Multiple agents coordinating in production, including handoffs, shared state, failure propagation.
• Large-scale data ingestion, or integrations with ads, social, CRM, analytics, or payment platforms.
• Open-source work, writing, or talks on agent systems. Experience with MCP. Fine-tuning or distilling smaller models to cut cost.

Our Stack

Python and TypeScript. LangChain and LangSmith for orchestration and tracing. MongoDB and ClickHouse for data. AWS for infrastructure. Deep integrations with Google/Meta Ads and other marketing platforms. You do not need to know all of it. You need to be able to pick it up fast.

Why Join Us

You will own the product and the technology behind it, and you will see the results in our customers' businesses. This is not a chatbot bolted onto an existing product. Agents, context, tools, data, and integrations are the core architecture, and the architecture you design becomes the platform's.

We grew out of content creation, so we move fast, experiment a lot, and care about what we make. We are backed by top Silicon Valley investors, serve customers around the world, and build from Singapore. Small senior team, real autonomy, no layers between you and the decision. If you want to build something ambitious with people who are all in, this is the place.

We care about output, not hours. Nobody tracks when you arrive or leave. We do expect real ownership: time spent learning, and what you learn brought back to the team. In return you get flexibility and the support to do your best work. Work always costs you time. We want that time to come back to you as something you are proud of.

What We Offer

• Competitive package with stock options.
• Hybrid work in Singapore. Employment Pass sponsorship and relocation support.
• Self-care and mental-health funds.
• Company lunches and a monthly SGD 250 Grab allowance.
• Yearly company retreats.

If you want to build AI that runs real businesses, and see the results, we would love to hear from you.