You'll help build Project Animus — our internal AI agent for revenue and customer intelligence. Your work turns scattered GTM data (call transcripts, CRM records, support tickets, telemetry) into clear answers and usable workflows for Sales, Success, Product, and Marketing.
This role is for someone who has genuinely internalized AI as a force multiplier — not just as a product to build, but as the primary way they work. You direct agents, review their output critically, run multi-step sub-agent review pipelines, and embed tools like Claude Code, Cursor, or similar agent tooling throughout your development lifecycle (AI-DLC — AI-native Development Lifecycle). You build faster and at a higher quality because AI is part of every step of your process.
You'll work closely with our AI Solutions lead, implementing and iterating on AI workflows on a Python/FastAPI + AWS foundation, partnering with GTM stakeholders to test and refine, and helping keep the system accurate, observable, and cost-aware.
What you’ll do here
Implement and iterate on AI workflows on top of existing data pipelines — product feedback extraction, customer update generation, onboarding plans, win/loss summaries, CRM enrichment
Extend and improve the FastAPI + LangGraph agentic loop: tool definitions, routing logic, prompt strategy, error handling, and observability
Integrate new data sources (Zendesk, Google Drive, telemetry, email) into the existing AWS Lambda pipeline architecture
Define and consume pre-aggregated account/opportunity summaries in S3 for fast, reliable structured query
Optimize Lambda-based data processing jobs for cost, reliability, and performance
Iterate on model strategy: cost-efficient routing (e.g., Claude Haiku) vs. higher-quality responses (e.g., Claude Sonnet/Opus) — knowing when to spend tokens and when not to
Evaluate prompts, tool selection quality, and response accuracy with clear metrics; build toward measurable quality in an agentic system
Collaborate with GTM stakeholders (Sales, SE, CS, Product, Marketing) to define, test, and refine AI-assisted workflows
Contribute to the web UI layer (React/TypeScript) as needed — leveraging agent tooling to execute frontend work efficiently
Contribute to the long-term evolution of the integration layer, including emerging patterns like Salesforce MCP for structured CRM access
What we’re looking for
3+ years in software engineering, including 1+ years building LLM-based applications or agents
Genuine fluency with AI-native development (AI-DLC): using agents and sub-agents to write, review, and iterate on code as standard practice — not an experiment
Strong Python for backend and data processing; FastAPI or equivalent async framework preferred
Hands-on experience with LLM tool calling and agentic loops — LangGraph, LangChain, N8N, or equivalent
Practical RAG expertise: chunking, metadata schemas, retrieval quality, and evaluation
Cloud infrastructure experience is a strong plus — AWS (Lambda, S3, Bedrock) preferred, but comfort with Railway, GCP, or similar and solid CI/CD experience is equally valued
Background in data integration and pipelines: consuming APIs and webhooks (Salesforce, Zoom, Slack, Zendesk, etc.)
Skill in designing JSON schemas, pre-aggregated summaries, and metadata models for structured query
Familiarity with Salesforce data structures (accounts, opportunities, leads) or willingness to ramp quickly
Experience with prompt engineering and evaluation for business workflows — accuracy, reliability, and user trust in AI-generated outputs
Comfort with modern JavaScript frameworks (React/TypeScript preferred) for contributing to frontend work
Openness to figuring things out together — we are actively learning how to integrate fully agentic workflows into a team environment and need someone who embraces that process, not someone who needs it fully defined before they start
Ability to travel up to 20%
Bonus tacos if you have
(Tacos? If you need an ice-breaker, ask how we say thanks by giving tacos!)
Amazon Bedrock knowledge bases.
Terraform and/or Kubernetes.
Slack bot development and slash commands.
CRM enrichment, sales tooling, or GTM analytics.
GDPR/data privacy considerations for AI systems.
Prior work on sales/revops intelligence tools, conversation intelligence (Zoom/Granola/Gong‑style), or human‑in‑the‑loop review flows.
Coder is an AI software development company leading the future of autonomous coding. We empower teams to build software faster, more securely, and at scale through the collaboration of AI coding agents and human developers. Our mission is to make agentic AI a safe, trusted, and integral part of every software development lifecycle.
Our self-hosted AI Development Environment is the foundation for deploying agentic AI in the enterprise. It provides a secure, standardized, and governed workspace to deploy autonomous coding agents alongside human developers, accelerating innovation while maintaining control and compliance. Coder's isolated, policy-driven environments improve productivity, cut cloud costs, and reduce data risks. Developers transition to AI at their own pace using their own tools. Platform and security teams can govern, audit, and manage a great developer experience at scale.
We believe that the interview process should be transparent, consistent, and enjoyable. We value your time and hope to complete the interview process in two to four weeks, if schedules allow. Through your interviews, you will meet a mix of individual contributors, managers, and senior leaders.
As an AI company, Coder embraces the use of AI tools, and we want to be transparent about our expectations as you navigate our interview process.
Not permitted: Using AI assistance during conversational interviews.
Permitted: Using AI tooling for take-home assessments. Please flag where and to what extent it was used in your take-home. Your submission will not be penalized for using AI as long as it is done honestly.
We use AI tools to help manage our recruitment process efficiently and fairly. Specifically:
Ashby helps us review inbound applications by surfacing candidates who best match the role requirements we've defined. This tool does not make hiring decisions - it helps our team prioritize which applications to review first.
Granola takes notes during our interview calls so our team can focus on the conversation with you.
All hiring decisions are made by humans. Our team reviews applications, conducts interviews, and makes final selections. AI tools assist us but never replace human judgment, and these practices are conducted in compliance with applicable data protection, AI governance, and labor laws. Your data is not used to train AI models.
In accordance with New York City Local Law 144, an independent bias audit has been conducted on "Automated Employment Decision Tools"; results are available for Ashby.
If you're applying for a role at Coder and have questions about how we use AI in our process, or if you'd like to request information about the data we collect, please contact careers@coder.com.
We are committed to providing equal employment opportunities to qualified applicants and do not discriminate on the basis of race, color, ancestry, religion, sex, pregnancy, gender, gender identity, gender expression, sexual orientation, national origin, age, marital status, genetic information, disability, protected veteran status or any other characteristic protected by federal, state, or local laws.
Want to know if this job is worth applying to?
Onsite