Associate, AI Data Scientist, Private Markets

About This Role

Build efficient, trusted agentic AI for private-markets intelligence.

Key Responsibilities

• Work forward-deployed with researchers, data specialists, and product teams to understand workflows, frame ambiguous problems, define measurable outcomes, and iterate rapidly from evidence to production.
• Research, design, and deliver agentic workflows spanning planning, retrieval, tool use, structured extraction, reasoning, verification, exception handling, and human escalation, using single-agent or multi-agent patterns where demonstrably appropriate.
• Engineer advanced agent patterns, including bounded iterative loops, reflection and critique, planner-executor, and evaluator-optimizer patterns, with explicit stop conditions, budgets, and safeguards.
• Build agent and model harnesses that standardize tool interfaces, context assembly, state and memory handling, policy enforcement, tracing, replay, and failure recovery.
• Determine the best fit for an agentic architecture, retrieval workflow, specialized model, or deterministic service, and validate against representative tasks and simpler baselines.
• Design safe caching and reuse across retrieval, prompts, model responses, tool results, and intermediate agent state, with clear cache keys, invalidation, freshness, access-control, and provenance requirements.
• Build systematic evaluation harnesses using representative datasets, explicit quality dimensions, adversarial and edge-case testing, regression suites, human feedback, and monitored production outcomes.
• Research and benchmark open-source SLMs and LLMs for task fit, considering quality, inference performance, cost, licensing, security, deployment constraints, and operational supportability; apply fit-for-purpose routing and batching where they improve efficiency.
• Develop model-adaptation strategies using prompt and context optimization, supervised fine-tuning, parameter-efficient fine-tuning, or other suitable techniques, supported by governed dataset creation, experiment tracking, reproducibility, and controlled promotion criteria.
• Operate the lifecycle for trained and fine-tuned open-source models, including data and model versioning, lineage, experiment and artifact management, evaluation gates, registry and release controls, deployment, performance and drift monitoring, rollback, retraining triggers, and retirement.
• Build production-quality Python components that integrate models with enterprise data, retrieval services, APIs, workflow orchestration, and human-review experiences.
• Improve grounding and trust through attribution, provenance, uncertainty handling, validation checks, access controls, and clear separation of source evidence from model-generated conclusions.
• Own delivered reliability across testing, observability, latency, cost, and maintainability; diagnose model, data, prompt, and orchestration failures in production.
• Document findings, reusable patterns, and trade-offs; contribute to design reviews; and communicate results clearly to research, product, engineering, risk, and senior stakeholders.
• Partner with engineering, product, risk, privacy, information security, legal, and compliance colleagues to support responsible, secure, and accountable use of AI.

Requirements & Tech Stack

• Strong experience building and deploying AI or machine-learning systems in an enterprise, SaaS, data-product, or similarly complex environment, with evidence of taking work beyond prototypes.
• Practical depth in LLM applications and agentic systems, including retrieval, tool use, planning, orchestration, memory or state management, structured outputs, verification, and evaluation of non-deterministic behavior.
• Experience implementing and comparing bounded loops, reflection or critique, planner-executor, and evaluator-optimizer patterns against simpler baselines, with measurable treatment of quality, reliability, latency, token usage, cost, and failure modes.
• Experience building agent or model harnesses across tool interfaces, context and state management, orchestration, policy enforcement, testing, tracing, replay, deployment, and observability.
• Hands-on experience evaluating, selecting, training, or adapting open-source SLMs or LLMs, including benchmarking, governed dataset preparation, supervised or parameter-efficient fine-tuning, and inference optimization.
• Practical ModelOps experience for trained or fine-tuned models, including dataset and model lineage, experiment and artifact tracking, registries, evaluation gates, controlled release, performance and drift monitoring, rollback, retraining, and retirement.
• Experience designing model routing, batching, and safe caching with appropriate keying, invalidation, freshness, privacy, access-control, and provenance safeguards.
• Strong grounding in modern machine learning, experimentation, and statistical evaluation, with disciplined handling of train-test separation, leakage, uncertainty, reproducibility, and model comparison.
• Ability to write production-quality Python and work effectively across data preparation, model development, APIs, orchestration, testing, deployment, and observability.
• Experience with unstructured or multimodal information, semantic retrieval, provenance, and data-quality engineering in workflows where domain experts provide ground truth and structured feedback.
• Experience working directly with users or domain specialists to turn ambiguous needs into usable capabilities and iterate based on observed workflow outcomes.
• Sound judgment on model and architecture trade-offs across efficacy, efficiency, security, licensing, maintainability, and operational risk.
• Clear communication and collaborative delivery skills, with the ability to explain experiments, limitations, and production trade-offs to technical and non-technical audiences.
• Commitment to responsible AI, data provenance, privacy, security, and meaningful human control in high-trust enterprise workflows.

Qualifications

• Understanding of private markets, alternative investments, financial data, or adjacent institutional-investment workflows is desirable; curiosity and the ability to build domain depth quickly are essential.