Founding Applied AI Engineer - Data
About the job
- Strong experience across some combination of data science, data engineering, and machine learning.
- Ability to design systems that reconcile real-time streaming data and batch or warehouse data under real latency constraints.
- Fluency in reading and questioning machine learning evaluation metrics, including AUC, precision and recall, and confusion matrices, to assess whether models will hold up in production.
- Demonstrated data skepticism, including questioning whether apparent results are accurate and examining raw records rather than relying on aggregates.
- Fluency using AI-assisted coding tools as a primary workflow, with the judgment to verify and defend analysis independently.
- Product instinct and motivation to build tools that make data work easier and more repeatable.
- Understanding of what modern AI, machine learning, and large language model systems need from data, including features, retrieval, context, and embeddings.
- Ability to take ownership and communicate effectively while working directly with customer teams.
- Turn high-value use cases into production by working directly with customer operators and Percepta engineers to drive transformation through data and AI.
- Own the full lifecycle of production models, including featurizing data, standing up serving pipelines, and building retraining and monitoring loops.
- Build end-to-end pipelines across streaming and batch sources to convert fragmented enterprise data into AI-ready assets.
- Integrate large language models into production pipelines, including structured extraction from unstructured text such as clinical notes, and own prompt quality and output reliability evaluation.
- Build internal products and tooling that make data work faster and repeatable across customers.
- Form and communicate technical opinions on data models, storage, orchestration, and infrastructure tradeoffs.
- Experience building agentic or automated data-engineering tooling.
- Hands-on experience with modern cloud data platforms such as Databricks and streaming systems such as Kafka.
- Experience with health-system data, including electronic health records, claims, admission-discharge-transfer feeds, and other operational healthcare datasets, or other complex regulated enterprise data.
- Experience using a large language model as a structured-extraction step inside a larger data pipeline rather than only as a chat interface.
- Prior startup, founding, or forward-deployed experience.
About the company
Percepta Labs builds cloud-based video analytics for retail loss prevention. It ingests retailers’ existing security camera feeds and uses action recognition and pose estimation to identify behaviors that could indicate shoplifting in real time, then assigns a probability that a theft is happening. The system emphasizes ethical AI by anonymizing demographic features (race, gender, age) to avoid bias and protect privacy. The product is sold to retailers to help reduce shrinkage and improve margins, with early testing claiming higher accuracy than human loss-prevention staff. Percepta’s growth has been supported by a major investment from ADT Commercial, which acquired the company and appointed its co-founder as ADT’s first Director of Artificial Intelligence. The company’s goal is to provide scalable, privacy-conscious video analytics that help retailers prevent theft more efficiently and cost-effectively.
Company Size
1-10
Company Stage
Seed
Total Funding
$1M
Headquarters
New York City, New York
Founded
2019
Get referred to Percepta
See people who can refer or advise you
Simplify's Take
What believers are saying
- Janus Henderson's June 2026 rollout spans 480 billion dollars in assets and 75 million clients.
- Percepta's Mosaic platform now powers client engagement and research workflows at Janus Henderson.
- Anthropic partnership gives Percepta credibility selling frontier AI inside regulated enterprises.
What critics are saying
- Palantir's January 2026 lawsuit alleges trade-secret theft and employee poaching.
- A Manhattan judge barred Percepta founders from soliciting Palantir workers in February 2026.
- Percepta depends on a few headline clients; losing Janus Henderson would crush growth.
What makes Percepta unique
- General Catalyst owns Percepta, giving it capital, enterprise access, and transformation expertise.
- Percepta embeds engineers inside clients, unlike software vendors selling off-the-shelf tools.
- Janus Henderson chose Percepta in June 2026 for PRISM and LIBROS infrastructure.
Help us improve and share your feedback! Did you find this helpful?
Benefits
- Health Insurance
- Dental Insurance
- Vision Insurance
- Life Insurance
- Disability Insurance
- Health Savings Account/Flexible Spending Account
- Unlimited Paid Time Off
- Flexible Work Hours
- Remote Work Options
- Paid Vacation
- Paid Sick Leave
- Paid Holidays