Lead AI Engineer, Platform

Posted 9 Hours Ago
3 Locations
In-Office or Remote
Expert/Leader
Software
The Role
Lead the technical direction for an AI product team building production language-model features, predictive models, agents, and an MCP server. Design retrieval, prompts, evaluations, release gates, monitoring, and production operations for healthcare software handling regulated data. Write and review primarily Node.js and TypeScript code, train predictive models, manage model selection and costs, enforce data permissions, mentor engineers, and communicate AI capabilities to product and business stakeholders.
Summary Generated by Built In

ECP is a market-leading SaaS solution that enables senior living communities to better care for their residents. ECP is used in over 8,000 communities. We're looking to further expand by increasing the number of customers that use our software and increasing the scope of how we serve our customers by developing and releasing new products. 

Senior living is deeply under-penetrated with software and ECP is one of the largest and fastest-growing software companies in the industry. We recently raised a growth round of equity to reinvest in our product, technology, and go-to-market. Our mission is to build world-class software that improves the quality of life for seniors and improves clinical, business, compliance, and operational performance for our customers. 

The Role 

ECP Intelligence is the AI in our products. Customers want more of it, so we're forming a team that will work on it exclusively. The team is two people, you and a senior AI engineer, both reporting to the VP of Engineering for Platform. 

You lead the technical work and decide how features are designed, evaluated and released. You and your teammate share the feature work, and you own the team's plan. Most of your time goes to writing and reviewing code. 

The features appear inside our clinical, billing and CRM products, so you'll work with the teams that own those screens. You'll also work closely with the product manager, who agrees what counts as good enough for release. Caregivers act on what these features tell them, so before a feature ships we need to be confident on the quality of the output, and we need to keep measuring that after it's live. 

The stack. We write code primarily in Node and TypeScript, with some Python for data-related work. Operational data is in SQL Server and PostgreSQL, and we have a data warehouse on Databricks. ECP is a HIPAA product, so resident data stays inside our AWS account, and models run on Amazon Bedrock or we host them ourselves. 

What You'll Build 

Features on language models. These include summaries of a resident's recent history and of a shift, speech to text for charting, and document intake that reads a scanned or faxed document and files it to the right resident once a person confirms it. For each feature you'll choose the model, design the retrieval and prompts, and manage cost and response time. 

Predictive models. These flag residents at rising risk of a fall or a change of condition. This is classical machine learning on tabular and time-series data, and we train these models ourselves. 

Agents and an MCP server. This team will add an MCP server that lets a customer reach their ECP data from their own AI tools, and agents that take actions in the product. An agent has to act as a specific user, inside that user's permissions. 

Evaluation and release. Every feature gets an evaluation plan before it leaves prototype, test sets drawn from real usage, a gate that every model or prompt change has to pass, and quality monitoring once it's live. We have a draft of this process. You'll own it and write the tooling for it. 

Production. The team runs its own features in production. That includes alerts and the ability to turn a feature off or roll it back. We don't build language, OCR or speech models from scratch. We start from ones we can run inside our AWS account, and you'll choose which for each feature and decide when one needs tuning. You'll also decide which models and tools are allowed to handle resident data, based on where the data is sent and what the vendor's contract permits. 

What We're Looking For 

Required 

  • You've shipped features on language models to paying customers and run them in production afterward. You know what they cost to run and how they failed. 
  • You've written the evaluation for an AI feature and checked its scoring against human judgment. You used it to decide on releases, and it stopped at least one change from shipping. 
  • You've trained a predictive model on tabular or time-series data, shipped it, and measured it against real outcomes after launch. 
  • You've put sensitive or regulated data through a model, and you can say where the data went and which terms covered it. 
  • You've led a team's technical work without managing the team. You set the direction, reviewed designs and code, and mentored the other engineers. 
  • You're a strong software engineer apart from the AI. You've designed services and data models, tested them, and been on call for them. Most of our code is Node and TypeScript, and you've worked in them or in a similar stack. 
  • AI-augmented development is how you work. You run several agents on a task or several tasks in parallel, with one agent's output reviewed by another before you review it yourself. You've set up the shared instructions and review rules a team uses for this. 
  • You're a strong communicator. You've written the design documents a team worked from, and you've explained what an AI feature can and can't do to product managers, executives and customers. 

Preferred 

  • Healthcare, or time spent watching clinical or frontline staff work 
  • Agents that acted for a user, where you decided what they could reach and enforced it in code 
  • An MCP server, or another tool surface that agents outside your company used 
  • Document extraction with OCR, or speech recognition, in production 
  • Amazon Bedrock, or open models you hosted yourself 
  • NodeJS, Python, SQL Server, Databricks 

What Success Looks Like 

At six months, several ECP Intelligence features are generally available, and customers are paying for them. Each one has an evaluation plan, a release gate, and quality monitoring in production. 

After that the work moves toward agents. We'll judge it by how much time our customers' staff get back. 

Skills Required

  • Experience shipping language-model features to paying customers and operating them in production
  • Experience creating AI feature evaluations, validating scores against human judgment, and using them for release decisions
  • Experience training and launching predictive models on tabular or time-series data and measuring real-world outcomes
  • Experience processing sensitive or regulated data through models and understanding data handling locations and contractual terms
  • Experience leading technical work without direct people management, including setting direction, reviewing designs and code, and mentoring engineers
  • Strong software engineering experience designing services and data models, testing systems, and supporting on-call production services
  • Experience with Node.js, TypeScript, or a similar software stack
  • Experience using AI-augmented development workflows with multiple agents, shared instructions, and review rules
  • Strong written and verbal communication, including technical design documents and explaining AI capabilities to stakeholders
  • Healthcare experience or experience observing clinical or frontline staff
  • Experience building agents that act for users with enforced permissions
  • Experience building an MCP server or another external agent tool surface
  • Production experience with document extraction using OCR or speech recognition
  • Experience with Amazon Bedrock or self-hosted open models
  • Experience with Python, SQL Server, or Databricks
Am I A Good Fit?
beta
Get Personalized Job Insights.
Our AI-powered fit analysis compares your resume with a job listing so you know if your skills & experience align.

The Company
HQ: Woodridge, IL
63 Employees
Year Founded: 2004

What We Do

ECP is a cloud-based software program for the senior living industry. The program has many unique features including electronic medication administration record (eMAR) that interfaces with pharmacies to sync med orders and reorder meds. It also has a complete care management program including integrated assessments, care plan wizard, point of care charting, alerts and reminders including emails and text messages, state mandated reports, marketing and billing. ECP is touch screen friendly and can be used on PC, Mac, netbooks, tablet, and iPad operating systems. ECP supports thousands of communities world-wide. ECP easily interfaces with LTC pharmacies. Resident information and medication orders are synced with the pharmacy improving communication and reducing medication administration errors. ECP is priced as a SAAS (software as a service) model reducing start-up costs for senior living communities. Users pay only for their current active occupancy level. ECP's customized settings meet the unique need of each company without custom programming. ECP allows quick creation of custom assessments and care plan libraries that automatically generate daily tasks for point of care charting.

Similar Jobs

Tie Logo Tie

Platform Engineer

Artificial Intelligence • eCommerce • Marketing Tech • Analytics
Remote
United States
54 Employees
135K-135K Annually

STN Inc Logo STN Inc

Software Engineer

Cloud • Information Technology • Cybersecurity • Infrastructure as a Service (IaaS)
Remote
USA
99 Employees

Root Logo Root

Lead Software Engineer

Automotive • Insurance • Machine Learning • Mobile • Software
Remote
United States
750 Employees
164K-240K Annually

MetLife Logo MetLife

Data & Operational Insight Analyst - 21457

Fintech • Information Technology • Insurance • Financial Services • Big Data Analytics
Remote or Hybrid
United States
43000 Employees
71K-120K Annually

Similar Companies Hiring

Ford Energy Thumbnail
Automotive • Software • Energy • Utilities • Manufacturing • Renewable Energy
US
Revel Thumbnail
Aerospace • Hardware • Robotics • Software
Marina Del Rey, California
70 Employees
Blee Thumbnail
Artificial Intelligence • Marketing Tech • Software
New York, New York
33 Employees

Sign up now Access later

Create Free Account

Please log in or sign up to report this job.

Create Free Account