$190K–$235K base salary + competitive equity. Relocation support available ($5K).
We are looking for a Forward Deployed Engineer with 3+ years of experience to own enterprise customer relationships end-to-end — from pre-sales demos to post-sales integration — for Honcho, our cutting-edge AI memory and identity infrastructure. You'll be the first dedicated FDE at a fast-growing neolab (50x developer growth in 3 months) that trains its own models and is the #1 trusted memory provider in the agentic AI ecosystem.
What will you be doing?
Managing deal flow end-to-end: triaging enterprise inbound, nudging customers through the pipeline, and delivering on commitments
Building demos and collateral (slide decks, one-pagers, reference integrations) tailored to each customer's vertical and use case
Embedding with customers post-sale — auditing their codebases and building Honcho integrations directly
Educating enterprise customers on agentic engineering mental models and how to leverage Honcho
Feeding customer friction and insights back to the dev team to improve the product
Plastic Labs is an engineering-driven AI lab building at the intersection of machine learning and cognitive science. The company is solving personal identity for the agentic world by developing Honcho — continual learning infrastructure for AI memory and identity modeling.
Honcho serves as the personal identity layer for AI-powered applications and agents. It uses theory-of-mind inference to synthesize high-fidelity representations of users, agents, and entities, enabling bottom-up alignment and deeply personalized AI experiences. We train proprietary language models to perform fast, cost-effective identity modeling — reasoning against all available data to create formal logic trees that help agents steer their behavior in real time.
Originally founded as an AI tutoring company, Plastic Labs pivoted within months to a horizontal infrastructure product after recognizing that every AI application builder faces the same fundamental challenge: they need rich, portable context about their users. The team went through the Betaworks accelerator program in New York and published research validating their methods.
Plastic Labs is backed by Long Journey Ventures, Four Rivers Group, Mozilla Ventures, Greycroft, Betaworks, Differential Ventures, Variant, and notable angels including the founders of Weights & Biases and Arcee AI, as well as early engineers at Facebook. The company is committed to open source AI, user-centric data practices, and a radically decentralized approach to alignment.
Team size
15 employeesFounded
Website
plasticlabs.aiTotal funding
Company locations
Plastic Labs is an engineering-driven AI lab building at the intersection of machine learning and cognitive science. The team of ~11 people operates out of their Williamsburg office in the Domino Refinery, Brooklyn, working five days a week in-person with a high-touch, collaborative culture.
The company was founded in 2023 by three co-founders — including a CTO with a deeply generalist background spanning hardware, cloud security, pen testing, Tesla Autopilot, PM work, and frontend engineering. The other two co-founders bring extensive experience in Machine Learning and Cognitive Science. Together, they pivoted from an AI tutoring product into Honcho, an open-source infrastructure layer for AI memory and identity modeling.
The engineering team currently consists of four engineers plus the CTO, with additional interns and contractors. Everyone reports directly to the CTO in a flat structure. The culture prizes high-agency autodidacts — people who thrive with broad freedom, significant responsibility, and minimal structure. The team values intellectual diversity, interdisciplinary curiosity (cognitive science, linguistics, neuroscience, philosophy), and a techno-optimistic mindset.
Day-to-day, engineers operate much like founding engineers — owning end-to-end projects, contributing to open-source community management, engaging with users in Discord, and acting as thought partners on product direction. The team builds in the open, teaches each other daily, and moves fast with a target cadence of releases every two weeks.
What defines success here:
Self-directed problem solving without needing explicit instructions
Genuine curiosity about AI and willingness to experiment with new tools
Generalist mindset — everyone wears multiple hats
Comfort with ambiguity and rapid iteration at the edge of emerging tech
Honcho is the personal identity layer for AI-powered applications and agents. It provides continual learning infrastructure for AI memory and identity modeling, enabling developers building AI applications to store and reason about context for their users, agents, APIs, codebases, companies, and teams.
Honcho takes a maximalist, reasoning-based approach — training proprietary language models to perform fast, cost-effective identity modeling. It reasons against all available data, enriching it to create formal logic trees of entities within the system. These representations help agents steer their behavior just-in-time to be more personalized and aligned with user intent.
Key capabilities include:
Theory-of-mind inference to synthesize high-fidelity user representations
Context management that eliminates cold-start problems for AI applications
Portable user data across applications — context travels with the user
Infrastructure-layer integration so builders don't have to reinvent memory and personalization from scratch
Honcho is open source and serves as a horizontal infrastructure product for the broader AI ecosystem. Developers building on top of Honcho can deliver deeply personalized experiences without building bespoke context management systems themselves.