Research Engineer, Agents at Equall in New York, NY

Why Equall

Equall is building the legal infrastructure for the private markets — a new system for how corporate legal reality is verified, tracked, and continuously managed.

Our AI-native platform extracts, structures, and reconciles corporate legal information, producing auditable outputs such as a cap table tie-out reports, governance dashboards, and personnel summaries grounded in source documents. Equall executes workflows end to end across both due diligence and day-to-day advisory work, verifying a company's legal state and surfacing actionable issues in real time.

In practice, this means:

• Investor-side due diligence delivered in hours, not weeks
• Company-side client snapshots, continuously updated
• Legal team review reduced by an order of magnitude

This opportunity comes at a pivotal moment in Equall's journey. We are working with leading law firms in ECVC and are now entering a period of rapid growth as we scale both our customer base and our product — extending coverage across the venture and emerging companies ecosystem and expanding into private equity and M&A.

Equall brings together leaders across legal, engineering, and AI to deliver transformative value. We listen closely to our customers, build with purpose, and work with both enthusiasm and a steadfast commitment to excellence — driven to reshape how corporate legal work is executed and how legal ground truth is determined.

If you are interested in playing an impactful role in this mission, we would be excited to hear from you.

The Role

We are seeking an exceptional Research Engineer to push forward the agentic systems at the core of Equall. Our platform reformulates legal work as a set of verifiable tasks — with structured outputs, objective correctness conditions, and a feedback loop that turns every user interaction into learning signal. You will design the agentic systems that ingest, extract, reconcile, and reason over legal state; build the benchmarks and evaluation infrastructure that make progress measurable; and advance the research that lets the platform learn from every workflow it runs.

What You'll Do

• Design, build, and evaluate agentic systems that reconstruct legal state end-to-end — document normalization, extraction, entity resolution, ambiguity handling, and validation against legal invariants
• Build our proprietary evaluation benchmarks and the Generative Dataroom Engine — synthetic data rooms with known legal flaws used to stress-test the system and benchmark releases
• Advance the platform's core learning loops: the inner loop that refines a company's graph from user interactions via semantic backpropagation, and the outer loop where the system evolves its own ontology based on observed friction
• Push toward agentic introspection — systems that propose and test structural changes to the schema itself based on usage patterns
• Partner with Legal Engineers and product to turn fuzzy legal workflows into verifiable tasks with crisp success criteria
• Contribute to and publish research — our team has a track record of publishing in the space

Our Research

Our team regularly publishes foundational work on AI for law. A few recent papers to get a sense of how we think:

• Does It Tie Out? Towards Autonomous Legal Agents in Venture Capital (http://arxiv.org/abs/2512.18658)
• SaulLM-54B & SaulLM-141B: Scaling Up Domain Adaptation for the Legal Domain (http://arxiv.org/abs/2407.19584)
• SaulLM-7B: A pioneering Large Language Model for Law (http://arxiv.org/abs/2403.03883)

What You'll Have

• Strong research and engineering fundamentals — you can read current ML papers in the morning and ship production code in the afternoon
• Hands-on experience building agentic systems — tool use, multi-step reasoning, planning, and the evaluation harnesses that keep them honest
• Experience designing benchmarks and evaluation infrastructure for AI systems, and a point of view on what "good" looks like
• Fluency in Python and the modern LLM stack
• Taste for problems where structure, verifiability, and iteration compound — where progress can actually be measured
• Curiosity about legal work and excitement about modeling corporate legal reality as a computable object

Bonus Points

• Prior research publications in NLP, ML, agents, knowledge representation, or programming languages
• Experience with RL, fine-tuning, or large-scale domain adaptation of LLMs
• Background in knowledge graphs, temporal data, type systems, or compilers
• Experience building synthetic data generation pipelines and data flywheels
• Experience with human-in-the-loop systems that turn user actions into learning signal
• Experience shipping agentic systems to production — not just prototypes