Client-Facing Developer

Perform Digital; United States; Remote; USD 110,000 - 170,000; Full time

Job summary

Perform Digital Pvt Ltd. in the United States seeks a developer who works directly with clients to build and operate agents in production and to communicate through scoping, training, support, and incident bridges.

The role requires 3+ years of full-stack development (Python or TypeScript) with at least 1 year using LLM APIs in production, plus exceptional English communication. You will own accounts end to end and collaborate with client teams on specs and go-live support.

Qualifications

• 3+ years of full-stack development (Python or TypeScript)
• At least 1 year using LLM APIs in production
• Excellent written and spoken English
• Familiar with agentic patterns: tool-use orchestration, RAG, multi-step planning, reflection loops
• Comfortable owning an account end to end and talking to humans about software

Responsibilities

• Build and deploy custom agents for assigned client accounts (2–4 accounts)
• Run discovery and scoping calls with client engineers, product managers, and ops leads; translate into a buildable spec
• Provide deskside support during go-live windows; be available on client Slack/Teams channels
• Own incident response for your accounts; triage, communicate, fix, and write the postmortem
• Train client teams on operating, tuning prompts, managing guardrails, dashboards, and extending agents

Skills

Full-stack dev

LLM APIs in prod

English communication

Account ownership

Empathy & patience

Tools

Slack

Teams

Job description

A developer who works directly with clients. Half the role is building and operating agents in production. The other half is communication: scoping, training, support, and the occasional incident bridge.

What you will do

• Build and deploy custom agents for assigned client accounts (between two and four accounts at a time)
• Run discovery and scoping calls with client engineers, product managers, and ops leads. Translate their language into a buildable spec.
• Provide deskside support during go-live windows. Be available on the client’s Slack or Teams channel.
• Own incident response for your accounts. Triage, communicate, fix, and write the postmortem.
• Train client teams on how to operate, tune prompts, manage guardrails, read dashboards, and extend the agent themselves

What we need

• 3+ years full-stack development (Python or TypeScript) with at least 1 year using LLM APIs in production
• Excellent written and spoken English. The other half of the job is communication.
• Familiar with at least two agentic patterns: tool-use orchestration, RAG, multi-step planning, reflection loops, structured-output extraction
• Comfortable owning an account end to end. You like to talk to humans about software.
• Empathetic and patient. You can explain a prompt regression to a non-technical PM without making them feel dumb.