Associate AI Data Scientist Private Markets

Skills

About the Role

You will research, build, and scale enterprise-grade agentic AI capabilities for private-markets workflows. You will design, evaluate, deploy, and monitor multi-agent and language-model systems, collaborate with users and technical stakeholders, and ensure reliable, secure, governed production outcomes.

Requirements

• Experience building and deploying AI or machine-learning systems beyond prototypes in complex environments.
• Practical experience with LLM applications and agentic systems, including retrieval, tool use, planning, orchestration, memory, structured outputs, verification, and evaluation.
• Experience implementing and comparing agent patterns against simpler baselines using quality, reliability, latency, cost, and failure metrics.
• Experience building agent or model harnesses across interfaces, state management, orchestration, testing, deployment, tracing, replay, and observability.
• Experience evaluating, selecting, training, or adapting open-source SLMs or LLMs.
• ModelOps experience including lineage, experiment tracking, registries, evaluation gates, release controls, monitoring, rollback, retraining, and retirement.
• Experience with model routing, batching, and safe caching.
• Strong knowledge of machine learning, experimentation, and statistical evaluation.
• Ability to write production-quality Python across data preparation, model development, APIs, orchestration, testing, deployment, and observability.
• Experience with unstructured or multimodal information, semantic retrieval, provenance, and data-quality engineering.
• Experience working with users or domain specialists to turn ambiguous needs into usable capabilities.
• Understanding of private markets, alternative investments, financial data, or adjacent institutional-investment workflows is desirable.
• Commitment to responsible AI, data provenance, privacy, security, and human control.

Responsibilities

• Work directly with researchers, data specialists, and product teams to understand workflows and define measurable outcomes.
• Research, design, and deliver single-agent and multi-agent workflows for planning, retrieval, tool use, extraction, reasoning, verification, and escalation.
• Engineer agent patterns including bounded loops, reflection, critique, planner-executor, and evaluator-optimizer approaches.
• Build agent and model harnesses for tool interfaces, context, state, memory, policy enforcement, tracing, replay, and recovery.
• Evaluate agentic architectures, retrieval workflows, specialised models, and deterministic services against representative tasks and baselines.
• Design safe caching and reuse for retrieval, prompts, model responses, tool results, and agent state.
• Build evaluation harnesses using representative datasets, adversarial testing, regression suites, human feedback, and production monitoring.
• Research and benchmark open-source SLMs and LLMs for task fit, performance, cost, licensing, security, and deployment constraints.
• Develop model-adaptation strategies, including prompt optimisation and fine-tuning.
• Manage the lifecycle of trained and fine-tuned models, including versioning, evaluation, release controls, monitoring, rollback, retraining, and retirement.
• Build production-quality Python components integrating models, enterprise data, retrieval services, APIs, orchestration, and human review.
• Improve grounding and trust through attribution, provenance, uncertainty handling, validation, and access controls.
• Own reliability across testing, observability, latency, cost, and maintainability.
• Document findings and communicate results to technical and non-technical stakeholders.
• Partner with engineering, product, risk, privacy, security, legal, and compliance colleagues on responsible AI use.

Benefits

• Retirement plan
• Tuition reimbursement
• Comprehensive healthcare
• Support for working parents
• Flexible Time Off