Member of Technical Staff (Data and Research)

Staying on top of thousands of stores, distribution centers, and replenishment managers is an age-old problem in consumer goods — one that is especially brutal for insurgent brands that are short on time and manpower. And it’s not a problem that will be solved with more software or more analysis: it’ll be solved with action and agents that can achieve real world results. That’s why Treater is the missing action engine for retail.

We're overturning a $200 billion industry of human sales reps that work for consumer brands from Olipop to P&G to ensure that stores are selling as much of their product as possible (typically through inefficient and lengthy store visits). Our AI agent replaces that manual labor by tirelessly working with store teams through email, text, and voice calling to ensure that every last revenue opportunity is captured for our customers.

Our mission is to enable brands and retailers to maximize sales while freeing them to focus on what they do best: delighting consumers. This will be a new world where, instead of relying on 4 disparate stakeholders in a 10-message email chain over two weeks to solve an issue that costs millions in sales, Treater's platform autonomously surfaces and resolves the same obstacle in hours. In just a few months, we’ve already driven millions of dollars in sales for a handful of launch partners.

We're looking for strong technical minds excited about this space to join us.

What you might work on

Please note that things are moving quickly. This list is representative of our needs today and may change tomorrow.

Generalized models to power our action engine
- Our long-term vision is to build a generalized engine that provides revenue opportunity recommendations to all actors in the retail space. Our existing platform is one input to this engine.
- Our existing models consist of regressive demand forecasting and out-of-stock anomaly detection. We're always looking to improve the quality of these models and build new ones for our customers.
- Fast-training and fast-inference models - we currently re-train our models daily or weekly, and perform inference on-demand.
- Opportunity to exhaust classical ML methods (we're particularly excited about sophisticated clustering and hierarchical modeling), and learning techniques.
Inference-time reasoning and recall for agentic systems
- Our platform consists of agents indistinguishable from humans that talk to stores in near real-time. There's always more work to be done to make communication a delightful experience for our users.
- Part of this relies on strong reasoning and memory systems, to perform relevant recall or even chat about last night's Lakers game. Systems of this kind are at the frontier of the space, and the general solutions are not sufficient for our needs; our specialized solutions are young but perform very well.
- Developing speech interruption models, systematically improving voice quality, automatically detecting resolving failure scenarios for our voice agents.
Data infrastructure
- Our product requires taking action on the most granular level of data, where volume and signal is often the lowest. You may work on methods to amplify signal and reduce noise, for example through hierarchical modeling.
- The data we ingest is often messy and inconsistent. You may work on methods to reconcile this data.
- Some of the data we ingest is partially anonymized (for example, store-level data does not have the store labeled). You may work on methods to de-anonymize this data.
- We are always looking for ways to improve our data pipelines, from ingestion to training to inference.

What we promise you

High autonomy, excellent cash and equity, and a lean team of people on your level. The founders were previously the first employees of startups and know what it takes.
Support from our incredible network of investors and advisors.
Work that you care about. As often as possible you will get your choice of what to do. The founders will take the rest.
Plenty of challenging work to be done, as described above.
You will work with everyone on the team and all of our customers. You will have a direct line to the founders. We do not build silos or org charts.
A strong technical team. Our team comes from Google, Lyft, D.E. Shaw, and many startups. We've worked on everything from web apps to distributed system infrastructure to commercial compilers/language tools to authentication systems to graph databases to container runtimes to real-time pricing systems.
Mentorship from within the company and our network.
Our job is foremost to deliver value to our customers, and secondmost to make you successful by metrics you believe in.

What we ask of you

Embrace a high-demand, high-autonomy role. We are looking to build a very lean team with high-agency individuals — so we can win as quickly as possible, with as little back-and-forth as possible.
Lead from the frontlines. You recognize output matters much more than meetings booked or lines of code written.
Passion for what we’re building and our customers. We will have tough days — and that’s when it’s really important for this job to be more than a salary or equity.
Work hard because you care — because you believe that everything we do should be excellent, and that you can’t stand missing a win because we weren’t willing to work just that extra little bit longer. The founders are also always on, and we promise that we’ll always be right there working alongside you.
Empathetic, thoughtful, and positive when collaborating with teammates or direct reports.
Pumped to work in person in New York City.
Appreciate what our customers do! CPG brands are really interesting to work with — and we’ll always have incredible snacks in our office.
You love writing code. You actively enjoy chatting about random technical topics.
You are excited about extracting signal from low-volume, messy data.
A strong foundation in, and intuition for, quickly building reliable data infrastructure. We also ask for understanding of both classical ML methods and standard learning techniques. Most of our data infrastructure is written in Python, with cleansed data stored in Postgres and raw data stored in unstructured sinks.

Why we're excited about this

It’s not every day that we get to transform one of the most influential and lucrative industries in the world into something that’s not just more efficient, but more democratic. Consumer goods should succeed because they delight consumers, not because someone remembered to email a distribution center a reminder at 11:17 pm the day before a new SKU launches. We take on the nitty-gritty so that the brands can get back to what they do best: bringing great products to the world.

Compensation

A competitive compensation package including base salary in the range of $120,000-200,000 and equity in the company. Final offers will be based on factors including experience and interview performance. We provide a benefits package with great medical, dental, and vision coverage. We have a 401(k) plan and provide PTO. As a principle, we believe in a small, tight-knit and very well compensated team.

Other perks

AI software subscriptions of your choice
We'll pay for any books or research materials that will help aid in your work.
Fully stocked office near Madison Square.

How to apply

Email [email protected] with your resume and a short introduction about yourself.