Staff Agentic Workflows Engineer, Enterprise Reasoning
New York, NY · Full-time · Hybrid
Applied AI
Hybrid
New York, NY
$235k – $325k
annual package
Role Dossier & Specification
REF: STAFF-AGENTIC-WORKTHE CLIENT & MANDATE
Our client is a multi-billion-dollar vertical AI leader transforming global law firms, corporate legal departments, private equity funds, and tax practices. From their New York and San Francisco engineering hubs, they build domain-specialized reasoning systems capable of executing multi-hour M&A due diligence, cross-jurisdictional regulatory analysis, and litigation drafting across thousands of interdependent source documents.
They have retained Talent247 to appoint a Staff Agentic Workflows Engineer for Enterprise Reasoning in New York, NY. Serving as an architectural lead on their core agentic platform, the successful candidate will design durable, multi-agent planning runtimes, verifiable citation engines, and adaptive context-compaction architectures operating at massive enterprise scale.
KEY RESPONSIBILITIES
- Architect durable, stateful multi-agent execution engines (built on Temporal, TypeScript, and Python) capable of decomposing complex legal and financial mandates into hundreds of parallel, fault-tolerant sub-tasks.
- Design dynamic context-management, hierarchical scratchpad memory, and cross-document synthesis patterns that prevent reasoning degradation across 1,000+ document data rooms.
- Engineer zero-hallucination span-level citation verification, contradiction detection, and structured tabular matrix generation pipelines audited by partner-level domain experts.
- Build multi-model orchestration layers that dynamically route planning, deep verification, high-speed extraction, and drafting steps across custom post-trained models and frontier reasoning APIs.
- Establish trace-level OpenTelemetry observability, replayable agent debugging sandboxes, and automated multi-turn evaluation gates catching latency or reasoning regressions prior to release.
- Mentor senior backend and applied AI engineers while setting technical standards across tool-interface contracts, tenant isolation, and execution reliability.
TECHNICAL ENVIRONMENT & OFFERING
Stack: TypeScript, Python, Go, Temporal, PostgreSQL, hybrid vector/lexical search, multi-provider frontier & custom post-trained LLMs, Kubernetes, Azure/AWS. Workplace: Hybrid in New York, NY. Compensation: $235,000 – $325,000 base salary plus substantial late-stage unicorn equity and comprehensive benefits.
Core Requirements
ESSENTIAL QUALIFICATIONS
- 7+ years of software and applied AI systems engineering experience, including Staff-level architectural ownership of high-scale distributed backend or agentic platforms.
- Deep production expertise building multi-step LLM agent orchestrators, durable workflow state machines (Temporal or equivalent), tool-use protocols, and structured output validation.
- Mastery of context engineering, hybrid retrieval orchestration, parallel map-reduce LLM execution patterns, and latency/cost optimization at high concurrency.
- Production fluency in TypeScript and Python (plus Go or Rust advantageous) with strong API design and multi-tenant security discipline (SOC2, tenant KMS encryption, zero data retention).
- Proven ability to partner closely with ML research teams to align runtime scaffolding with custom post-trained model capabilities.
PREFERRED / BONUS EXPERIENCE
- Prior experience building mission-critical AI workflows in legal tech, quantitative finance, M&A data rooms, or enterprise search.
- Hands-on experience designing evaluation frameworks for long-horizon, semi-verifiable multi-document agent trajectories.
- Contributions to open-source agent runtimes, MCP server ecosystems, or durable execution tooling.