Agentic Trader

About the role

We are opening an agentic trading seat: agents that watch markets, use tools, and propose or place trades only inside rules a human can stop. This is not a signal-chat role. An agentic trader owns the loop: observe the market, reason, act through tools, and learn from the result, with hard limits on size, loss, and exposure.

Key responsibilities

• Design trading agents that read market data, news, and portfolio state, then choose an action through tools.
• Backtest ideas with out-of-sample and walk-forward checks. Flag overfitting before anything goes live.
• Set risk limits: size, loss, exposure, and a kill switch. Keep a human approval step where it matters.
• Connect agents to market data and execution tools. Log every decision, tool call, and fill so a trade can be explained after the fact.
• Watch live runs, review failures, and improve the agent from real market feedback.
• Write short notes on what the agent did and why, for the rest of the team.

About you

• Hands-on with markets: crypto, FX, or equities. You can explain a fill, funding or borrow, slippage, and why a backtest lied.
• Python for research and production. Comfortable with APIs, time-series data, and basic statistics.
• Built or run LLM agents: tool use, memory, evals, and guardrails. A framework is optional. Owning the loop is not.
• Careful with risk. You would rather miss a trade than break a limit.
• Good teamwork and problem-solving. English is the main working language; spoken Thai is a plus.
• A shipped project, GitHub, or write-up of a strategy or agent is a plus.

About us

We are a Thailand-based startup building last-mile autonomous delivery robots. Our team designs, builds, and deploys robots for real delivery work, and we are the only Thai team selected for Protocol Camp Cohort 10, the pan-Asia builder program hosted by ShardLab and Hashed.

Additional details

• Working Hours: Normal business hours
• Location: The Urban Office Asoke, https://maps.app.goo.gl/SH3izYCvFqERm8CX9

Application

• Please send your Resume/CV (English preferred) and a short note on an agent or strategy you have built or run. A portfolio or GitHub helps.
• Email: [email protected]
• Contact: 02-200-1194