Staff AI Engineer

Relativity; Alton (IL); Hybride; USD 174 000 - 262 000; Plein temps

Résumé du poste

Relativity is seeking a hands-on Staff Engineer to own how Discovery systems are built, evaluated, and operated. You will shepherd end-to-end inference paths, from retrieval to outputs, and ensure robust observability in a multi-tenant cloud environment.

You will define evaluation metrics, optimize latency budgets, and collaborate across domains—engineering, product, and design—to deliver defensible, scalable solutions for legal data processing at scale.

Qualifications

• 7+ years building and operating large-scale distributed cloud systems.
• 3+ years shipping production model-backed systems.
• Depth across LLMs, RAG, agentic workflows, and vector/hybrid search.
• Owned an evaluation practice with eval sets, metrics and gates.
• Architecture accountability across cloud services in multi-tenant SaaS.
• Quantitative reasoning for model behavior including accuracy, latency, cost.
• Strong technical leadership and the ability to present tradeoffs to executives.

Responsabilités

• Staff engineers set technical direction across teams and build the hardest parts from prototype through production.
• Own the inference path end to end including retrieval, context construction, routing, and observability.
• Set the evaluation bar with gold sets, offline suites, CI gates, and online telemetry.
• Own token economics and latency budgets; decide routing, caching, and model tier based on data.
• Design for reuse by decoupling capability logic from delivery harness via interface contracts.
• Architect the data and retrieval layer including indexing, embedding, and hybrid search.
• Engineer defensibility with citation fidelity, confidentiality controls, and auditability.
• Lead root-cause analysis, publish learnings, and raise engineering standards.
• Collaborate across domains with Applied Science, Product, Design, and peers to drive closure.

Connaissances

Distributed systems

Production ML models

LLMs & RAG

Evaluation practice

Cloud architecture

Quantitative reasoning

Technical leadership

Description du poste

Posting Type

Hybrid/Remote

Job Overview

Who We Are Relativity is a leading legal data intelligence company building technology that helps users organize data, discover the truth, and act on it with confidence. Our AI‑powered, cloud platform, RelativityOne, transforms massive volumes of complex information into actionable insights for litigation, investigations, regulatory inquiries, data breach responses, and other high‑stakes legal work where accuracy and trust are crucial.

The world’s largest law firms, corporations and government agencies rely on Relativity’s legal AI software to securely surface and manage the most relevant and impactful information in their matters. Beyond our commercial impact, we’re deeply committed to expanding access to technology for academic institutions through Relativity Academic and supporting pro bono legal work through Justice for Change. What We Do At Relativity, engineers don’t just write code— they harness AI tools to architect, optimize, and deliver scalable software faster and shape the future of legal technology. They shape how industries uncover critical insights, ensure compliance, advance access to justice and continuously evolve our platform using the latest technologies, including cloud-native architectures, AI capabilities, and modern developer tooling. Our AI powers privacy programs, identifies confidential information, and detects language indicating misconduct. We’re reimagining how legal professionals uncover truth at scale, and engineering is at the heart of that transformation. If you’re passionate about solving hard problems, designing for scale, and building tools that matter—while working in a culture that values transparency, innovation, and shared success— Relativity offers the opportunity to make a measurable difference in the marketplace.

Job Description and Requirements

About the Role

Discovery builds the products at the center of Relativity: Document Review, aiR for Case Strategy, Intelligent Fact Gen, Memo Gen, Proof Charting, Translate, and Redact. They reason over petabyte-scale legal corpora, and their output must hold up in front of counsel, opposing counsel, and a court. What they surface and what they miss shapes the outcome of real cases.

You will own how these systems are built, evaluated, and operated. This is a hands‑on role across the whole Discovery portfolio. Not research, not management.

What You’ll Do

• Staff engineers here set technical direction across teams and build the hardest parts with them, from prototype through production. Knowing which designs to hand to a team and which to carry yourself is part of the job. You are not running a single team's backlog.
• Own the inference path end to end. Retrieval, context construction, model selection and routing, agent orchestration, tool interfaces, structured output, graceful degradation, and observability that makes non-deterministic failure debuggable at 2am.
• Set the evaluation bar. Golden sets, offline suites, regression gates in CI, online quality telemetry. A change to a prompt, a model, or a retrieval strategy gets measured, not argued.
• Own token economics and latency budgets. Instrument cost and latency per workflow, set unit targets, and decide routing, caching, and model tier from data.
• Design for reuse. Decouple capability logic from its delivery harness behind explicit interface contracts, so a single capability may be shipped as a service, an MCP skill, or an embedded experience without a rewrite.
• Architect the data and retrieval layer. Batch and streaming pipelines, indexing and embedding strategy, hybrid search, tenant isolation, and lineage that traces an answer back to the document it came from.
• Engineer for defensibility. Citation fidelity, privilege and confidentiality controls, human review paths, and the auditability a customer needs to defend a workflow under challenge.
• Catch drift before customers do. Quality regressions, hallucination and refusal rates, latency and cost creep. Lead root cause analysis, eliminate recurrence, and publish the learning across Discovery.
• Raise the bar across teams. Design review, reference implementations, documented paths from proof of concept to production, and mentorship of senior and lead engineers.
• Collaborate across domains. Work with Applied Science, Product, Design, and peer domains, and drive calls to closure where you hold no formal authority. Push coding agents and modern SDLC practice into the teams without lowering the engineering bar.

What We’re Looking For

Required

• 7+ years building and operating large-scale distributed cloud systems.
• 3+ years shipping production model-backed systems that real customers depend on. Prototypes and demos will not be enough.
• Depth across LLMs, RAG, agentic workflows, prompt and context engineering, vector and hybrid search, orchestration frameworks, and observability for non-deterministic systems.
• You have owned an evaluation practice. You built the eval sets, metrics, and gates that told a team whether a change helped or hurt.
• Architecture accountability across cloud services and data pipelines in multi-tenant SaaS, including scalability, reliability, and security.
• Quantitative reasoning for model behavior: eval metrics, error analysis, and tradeoffs across accuracy, latency, throughput, and cost.
• Technical leadership across teams without formal authority, and the writing skills to put a decision, its alternatives, and its tradeoffs in front of executives and get an answer.

Preferred

• Legal technology, eDiscovery, search, document intelligence, or another domain where output must be defensible and auditable.
• Agentic and multi-agent systems, including consolidating specialized agents into a generalist architecture without losing task quality.
• MCP servers or model-facing tools and skill interfaces.
• Cutting inference cost at scale through routing, distillation, caching, or smaller fine-tuned models.
• Governance, monitoring, and traceability practices in regulated or high-trust environments.
• Azure, Kubernetes, CosmosDB, and progressive delivery through feature flagging.

Relativity is committed to competitive, fair, and equitable compensation practices.

This position is eligible for total compensation which includes a competitive base salary, an annual performance bonus, and long-term incentives.

The expected salary range for this role is between following values:

$174,000 and $262,000

The final offered salary will be based on several factors, including but not limited to the candidate's depth of experience, skill set, qualifications, and internal pay equity. Hiring at the top end of the range would not be typical, to allow for future meaningful salary growth in this position.

Required Skills:

Algorithms, Automation, Debugging, Distributed Systems, Performance Tuning, Problem Solving, Project Management, Software Development, System Designs, Technical Leadership