> Markdown version of [/jobs/ext/4556062-ai-engineer](https://www.wearedevelopers.com/jobs/ext/4556062-ai-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# AI Engineer

- **Company:** Peloton Interactive, Inc.
- **Location:** New York, NY, United States
- **Experience:** Expert
- **Salary:** $160,600 - $224,950
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, BigQuery, Databases, Information Leak Prevention, Data Retention, Distributed Systems, Salesforce.Com, SAP (Applications), Software Engineering, Pinecone, Evaluation Pipelines, Retrieval-Augmented Generation, Delivery Pipeline, Large Language Models, Caching, Agentic-AI, Pgvector, Kubernetes, Low Latency, Entity Resolution, Microservices
- **Published:** November 4, 2026
- **Apply:** https://startup.jobs/sr-ai-engineer-full-stack-peloton-10660713

## About the Role

* 5-7+ Years of General Software Engineering. Strong foundational software engineering skills (distributed systems, async patterns, API design, microservices, and databases)
 * Comfortable pivoting to whatever language or framework gets the job done fastest
 * Hands-on experience deploying RAG (Retrieval-Augmented Generation) and Agentic Workflows into production using LLM APIs, vector databases, and orchestration frameworks
 * Proven track record building and maintaining pipelines across fragmented enterprise systems (CRM, ERP, CPQ, marketing platforms) with attention to entity resolution, schema design, and data quality - not just consuming clean data once it exists
 * Cutting-Edge Stack: Work at the bleeding edge of the AI ecosystem with top-tier tools and compute resources.
 * Direct Impact: What you build goes directly into the hands of real users in days, not months.

## Description

* Take ambiguous problems and turn them into functional, high-performing AI features and standalone tools
 * Solve the "State Problem" for AI by building memory architectures and vector stores (Pinecone, Pgvector) that maintain context while strictly adhering to data retention and deletion policies
 * Build pipelines and entity-resolution logic that connect data across CRM, ERP, CPQ, and marketing systems (e.g. Salesforce, SAP, BigQuery) into a single, reliable source of truth agents can act on
 * Embed with Peloton business units to understand their workflows, identify high-value opportunities, and select the right initial use cases to prove value
 * Embed closely with stakeholder teams, understand their needs, and guide adoption of what you build
 * Share field feedback that helps other teams understand where the models succeed and where they can improve
 * Deploy continuous evaluation pipelines to test models for hallucinations, prompt injections, data leakage, and performance regressions before code hits production
 * Implement Semantic Caching and smart routing logic (directing routine queries to smaller, cheaper models) to maintain low latency and efficient inference costs