A $40B long-only asset manager is hiring its first dedicated data science function. The COO must decide how to structure the team relative to existing investment and technology departments. Which organizational model best positions the data science function for long-term investment impact?

Organizational placement of the data science function determines whether it generates investment alpha or becomes an expensive internal IT team. The centralized data science center of excellence with embedded liaison roles that partner directly with investment teams, reporting through a Chief Data and AI Officer into the chief information officer's organization, signals that data science is a first-class investment capability, not a technology support function. This reporting line ensures that model prioritization aligns with investment strategy rather than IT roadmaps, and that data scientists are evaluated on investment-relevant KPIs such as signal quality, model-driven P&L attribution, and analyst productivity lift.

The centralized center of excellence model with embedded liaisons captures the best of both worlds. The center of excellence maintains shared compute infrastructure, model governance frameworks, data procurement relationships, and a common tooling stack -- preventing the duplication and technical debt that arise when every investment pod builds its own data science capability independently. Liaison roles -- data scientists who sit physically and organizationally adjacent to investment teams for rotating assignments -- ensure that model development stays grounded in actual investment workflows rather than becoming an academic exercise.

Placing the entire data science team within IT, reporting to the chief technology officer so that infrastructure and model deployment are handled within a single technical department, embeds data science in a cost-center mindset focused on uptime and infrastructure, not alpha. Embedding data scientists exclusively within individual investment pods so that each portfolio manager has dedicated quant support, with no centralized data science function, creates siloed, duplicated efforts, prevents knowledge sharing, and makes it difficult to recruit and retain strong data science talent who want to work with peers of comparable caliber. Outsourcing all data science work to a third-party analytics vendor and having investment analysts consume vendor-produced models through an API dashboard sacrifices proprietary data advantages and creates dependency risk; vendors cannot develop the firm-specific institutional knowledge needed for durable edge. Industry benchmarks from Man Group, Two Sigma, and Acadian Asset Management consistently show that firms with centralized data science organizations reporting into investment leadership outperform those where data science is siloed in technology.