Lead Machine Learning Engineer/Scientist - Algorithms & Research

Upwork Inc.'s (Nasdaq: UPWK) family of companies connects businesses with global, AI-enabled talent across every contingent work type including freelance, fractional, and payrolled. This portfolio includes the Upwork Marketplace, which connects businesses with on-demand access to highly skilled talent across the globe, and Lifted, which provides a purpose-built solution for enterprise organizations to source, contract, manage, and pay talent across the full spectrum of contingent work. From Fortune 100 enterprises to entrepreneurs, businesses rely on Upwork Inc. to find and hire expert talent, leverage AI-powered work solutions, and drive business transformation. With access to professionals spanning more than 10,000 skills across AI & machine learning, software development, sales & marketing, customer support, finance & accounting, and more, the Upwork family of companies enables businesses of all sizes to scale, innovate, and transform their workforces for the age of AI and beyond.

Since its founding, Upwork Inc. has facilitated more than $30 billion in total transactions and services as it fulfills its purpose to create opportunity in every era of work. Learn more about the Upwork Marketplace at Upwork.com and follow us on LinkedIn, Facebook, Instagram, TikTok, and X; and learn more about Lifted at Go-Lifted and follow on LinkedIn.

We’re looking for a Lead Machine Learning Engineer/Scientist. In this technical lead role, you will help build a Dynamic Memory Management capability for Upwork’s LLM-powered experiences, including agentic systems and tool-using assistants. This role sits at the intersection of retrieval, memory, reasoning, and orchestration, shaping how AI systems store, update, compress, retrieve, and apply knowledge across sessions, tasks, and workflows.

You will build production-grade memory architectures that combine structured and unstructured signals such as user preferences, entities, constraints, conversation history, tool results, marketplace context, and long-term facts. You will design memory policies, develop retrieval-augmented generation (RAG) and memory fusion strategies, and train or post-train models to execute tool calls grounded in memory and context. Success in this role is measured through clear Focals such as task success rate, factual consistency, hallucination reduction, latency, cost, and durable personalization impact.

Responsibilities

• Architect a Dynamic Memory Management system for LLM and agent applications, including memory ingestion, CRUD operations, retrieval, consolidation, summarization, and forgetting policies, meeting defined Focals for reliability and latency.
• Design RAG plus memory architectures that integrate vector databases, relational stores, and knowledge-graph representations to improve grounding quality and downstream task success.
• Develop multi-stage retrieval and ranking strategies, including re-ranking, salience scoring, recency versus importance tradeoffs, conflict resolution, and deduplication, validated via offline precision and recall and online outcomes.
• Build end-to-end pipelines for retrieval-augmented context flows, including modeling approaches for memory selection, compression, and safety-aware grounding.
• Train and post-train models for reliable function and tool calling, including tool selection, schema adherence, multi-step planning, and policy-compliant execution, with measurable reductions in invalid tool calls.
• Establish evaluation frameworks and monitoring across offline and online metrics such as memory precision and recall, factual consistency, hallucination risk, task success rate, latency, and cost, then translate insights into iterative improvements.
• Lead cross-functional delivery from prototype to production by partnering with engineering, product, and trust and safety on privacy boundaries, storage design, orchestration, observability, and incident-ready quality practices.

What it takes to catch our eye

• Demonstrated experience shipping LLM-powered agent or assistant systems to production with measurable impact on user outcomes, reliability, or cost.
• Strong depth in retrieval and ranking systems, including embedding strategies, hybrid retrieval, re-ranking, and evaluation methodologies for RAG and memory-aware applications.
• Hands-on expertise designing memory behaviors in LLM systems, including summarization, consolidation, forgetting, conflict resolution, and personalization policies.
• Proven ability to improve structured tool calling through post-training or constrained decoding methods, including dataset construction and quality gates for schema adherence and safety.
• Adaptive AI fluency in technical workflows: you use AI tools to accelerate experimentation, code and evaluation development, and debugging, and you review outputs rigorously for accuracy, risk, and production readiness while sharing best practices with teammates.

Come change how the world works.

This position will initially be employed through a partner to ensure a seamless hiring process while we establish the hub. Once the hub is established, there may be opportunities to transition to employment with Upwork depending on business needs and other requirements. While employed by the partner, you’ll work as part of Upwork’s team, with access to our resources, culture, and growth opportunities.

Please note that a criminal background check may be required once a conditional job offer is made. Qualified applicants with arrest or conviction records will be considered in accordance with applicable law, including the California Fair Chance Act and local Fair Chance ordinances. The Company is committed to conducting an individualized assessment and giving all individuals a fair opportunity to provide relevant information or context before making any final employment decision.

To learn more about how Upwork processes and protects your personal information as part of the application process, please review our Global Job Applicant Privacy Notice