Applied AI Engineer

Vation Ventures; Denver (CO); On-site; USD 150,000 - 210,000; Full time

Job summary

Vation Ventures seeks an experienced engineer to own the services and data layers underpinning AI applications for enterprise clients. You will ship an LLM-backed feature, model data in Aurora Postgres and DynamoDB, write Terraform, and build a frontend you can demo to a client CTO.

You’ll join a small team, own end-to-end delivery across backend, API, and frontend, and work directly with client teams on architecture and reviews.

Qualifications

• 5+ years shipping and maintaining production web applications on AWS.
• At least one LLM-backed feature deployed to production and supported afterward.
• Experience with Amazon Bedrock or equivalent hosted model platform.
• Strong TypeScript and Python for data and AI work.
• Experience with Next.js and React in production apps.

Responsibilities

• Develop AI services: inference, orchestration, tool calls across models.
• Design APIs with Lambda/API Gateway or ECS/Fargate.
• Model data in Aurora Postgres and DynamoDB with indexing.
• Write Terraform modules and manage CI/CD for infra.
• Collaborate with client teams on architecture and readouts.
• Frontend work with Next.js server components and streaming.

Skills

AWS

TypeScript

Python

Next.js

React

Aurora PostgreSQL

DynamoDB

Terraform

Serverless

API Design

CI/CD

Tools

Terraform

AWS Lambda

ECS/Fargate

API Gateway

pgvector

Amplify

S3

OIDC/SAML

Job description

We design and build production AI systems for enterprise clients. The work is applied rather than experimental: real data, real constraints, and a client team that will operate the system after we hand it over.

This role owns the services and data layers underneath those applications, along with the interface on top of them.

We're hiring for range. The engineer we're looking for has shipped an LLM-backed feature and supported it afterward, can model data in both Aurora Postgres and DynamoDB and explain why each one is there, writes their own Terraform, and can build a front end they'd be comfortable demoing to a client CTO. That's a broad profile, and we'd rather be clear about it now than halfway through an interview process.

What you'll work on

AI systems

• Inference and orchestration services on Amazon Bedrock: prompt and context assembly, tool and function calling, agentic loops, streaming, retries, and fallbacks across models.
• Retrieval, end-to-end. Embedding and ingestion pipelines, vector search in Aurora PostgreSQL with pgvector, hybrid vector and keyword strategies, chunking and re-ranking, and the query tuning that holds p95 steady as the corpus grows.
• The parts that determine quality: document parsing, extraction accuracy, PII handling, data contracts, and integration with the source systems clients already run.
• Instrumentation. Token and cost accounting, latency budgets, tracing through multi-step chains, and evaluation harnesses that show whether a change improved the system or only changed it.

Application and infrastructure

• API design and implementation: Lambda and API Gateway or containers on ECS/Fargate, auth and tenancy boundaries, async and queue-based work, and IAM scoped to pass a client security review.
• Data modeling driven by access patterns. Aurora Postgres for relational and vector workloads, DynamoDB where the access pattern calls for it. Schema design, migrations, indexing, and query plan analysis.
• Integrations with systems we don't control: Salesforce, ERPs, HRIS platforms, data warehouses, file transfers, and legacy SOAP endpoints.
• Infrastructure as code in Terraform: modules, environments, state, and CI-driven applies. Terraform owns the infrastructure, and Amplify is scoped to hosting, environments, and auth so the two don't overlap.
• Operations. Logging, alerting, performance budgets, and test coverage sufficient to deploy on a Friday.

Frontend

• Application interfaces in Next.js (App Router, TypeScript, server components and server actions) with production-grade forms, tables, state management, and error handling.
• The interaction problems specific to AI surfaces: streaming and partial output, latency that has to feel intentional, citations, and clear signaling when the system has low confidence.
• Hosting and operations on AWS Amplify: environments, auth flows, and preview deployments.
• Implementation against a design system. You don't need to be a designer, but you should be able to tell when something looks unfinished.

What the role offers

• Systems that go into production. Every engagement ends with a client team running what you built, not with a prototype that gets shelved.
• Ownership across the stack. You'll be one of two or three engineers on a build, with real authority over architecture rather than a narrow slice of someone else's design.
• Direct client access. You'll be in the room for architecture discussions and readouts instead of receiving requirements secondhand.
• Range across problem domains. Engagements vary by industry and system, so the technical problems don't repeat.
• Scoped work. Engagements are defined in a statement of work, with an end date and a defined deliverable.

Required skills

• 5+ years shipping and maintaining production web applications, with significant time on AWS.
• At least one LLM-backed feature taken to production and supported afterward. You know where RAG breaks, why agents stall, which evaluations are worth building, and what you'd do differently.
• Amazon Bedrock, or equivalent production experience with another hosted model platform and the judgment to transfer it.
• Strong TypeScript, plus Python for data and AI work.
• Next.js and React beyond marketing sites, with experience living alongside your own architecture decisions.
• Relational depth on Aurora or PostgreSQL: schema design, migrations, indexing, reading a query plan, and diagnosing slow queries methodically.
• DynamoDB experience where you drove the data model from access patterns, including single-table design and its tradeoffs.
• Terraform in practice: you've authored modules and untangled someone else's state.
• Serverless and container patterns on AWS: Lambda, API Gateway, ECS or Fargate, SQS, EventBridge, S3, and IAM you can explain in detail.
• Comfort working directly with client teams, including architecture discussions with their engineers and difficult questions in a readout.

Preferred skills

• pgvector at scale, or hybrid retrieval in production.
• Bedrock Knowledge Bases, Guardrails, or Agents.
• Document AI: OCR, layout-aware parsing, and table extraction from difficult PDFs.
• Enterprise authentication: SAML, OIDC, SCIM.
• Experience in regulated environments such as financial services, healthcare, or manufacturing.
• Consulting or client-services background, or small-team experience with end-to-end ownership.

This role may not be the right fit if

• You want to specialize in backend or frontend exclusively. Both are good careers, and this position is neither.
• You've moved into architecture and no longer write code regularly. This is a hands‑on role.
• Your AI work has been prototypes and demos without responsibility for a system in production.