Drimva

Invisible Technologies

Senior Solutions Architect - Data Labs

London - Hybrid +5 · Senior · Full-time · Posted Sep 25, 2026

Location
London - Hybrid
New York - Hybrid
San Francisco Bay Area - Hybrid
London, England, United Kingdom
New York, United States
San Francisco, California, United States
Education
A background in computer science, software engineering, machine learning, data science, AI research, technical product, solutions architecture, or a related field is welcome; equivalent knowledge through independent work or other professional experience is also welcome.

About the job

Leads the design of human-data programs and pilots for a frontier AI lab, translating client research goals into feasible, scalable training and evaluation solutions.

What you'll do

  • Own solution design from discovery and scoping through handoff and early delivery; clarify ambiguous client requests and align stakeholders.
  • Advise on data and evaluation methods, task and workflow design, quality strategies, expert profiles, and commercial feasibility.
  • Partner with Engineering on collection interfaces and technical architecture, test critical solution elements, and share reusable practices.

What they ask for

  • Strong working knowledge of LLM training and evaluation, including reasoning across modalities, collection methods, and research objectives.
  • Technical fluency in ML, data systems, software architecture, APIs, and data flows.
  • Exceptional communication, stakeholder influence, and judgment in ambiguous environments.
  • Commercial judgment for rough-cut budgeting and balancing research goals with expert availability, effort, timelines, costs, and unit economics.

Nice to have

  • Experience with human-data programs, annotation, post-training, RLHF, RLVR, evaluations, benchmarks, red teaming, or research operations.
  • Hands-on use of Python, SQL, APIs, notebooks, or lightweight prototyping.

Tools

Python, SQL, APIs

AI focus

Designs human-data programs for frontier AI training and evaluation, including supervised fine-tuning, preference data and RLHF, benchmarks, red teaming, multimodal data, coding tasks, and agent environments.

Model-written summary · the original posting is authoritative

First seen Oct 27, 2026 · Last verified Nov 5, 2026 (UTC)