Data Scientist - Portfolio Optimization

Formation Bio•Boston, MA

6d•Hybrid

About The Position

As a Data Scientist on the platform prediction team, you'll translate our probability of success predictions into measurable portfolio-level outcomes. You'll architect core systems — order management, execution simulation, portfolio construction, risk monitoring, and performance attribution — that let us rigorously evaluate signals from our AI-driven predictions in public and private equities and our internal portfolio. This role sits at the intersection of quantitative finance, healthcare data, and AI-driven drug development. If you're excited about applying portfolio construction and risk management fundamentals to one of the most consequential prediction problems in healthcare, this is the role.No other company — hedge fund or pharma — has a technical data science position translating drug development experience into durable AI-native portfolio strategies. The skills you develop here — portfolio construction over assets with radically asymmetric risk profiles, clinical trial analytics, AI/ML in production, and risk management across multi-year horizons — can directly impact the delivery of new and effective therapeutics to patients by best aligning impactful medicines with economic incentives.

Requirements

MS or PhD in a quantitative field (statistics, finance, physics, computational science, engineering, or related)
1-3 years in a quantitative research, data science, or analytics role — finance, healthcare, academic research, or consulting all count; substantive internships qualify
Strong Python programming skills with experience in data-intensive workflows (pandas, numpy, scipy)
Solid grasp of core portfolio construction and risk concepts: position sizing, rebalancing, Sharpe ratio, drawdown, volatility, benchmark comparison
Demonstrated ability to work with messy, real-world datasets — comfortable with data wrangling, deduplication, and quality assessment
Clear communicator who can present quantitative results to both technical peers and business stakeholders

Nice To Haves

Experience with backtesting frameworks or portfolio simulation (vectorbt, Backtrader, or custom implementations)
Exposure to healthcare, pharma, or biotech data (clinical trials, claims data, -omics, real-world evidence)
Familiarity with alternative data in a research or investment context
Experience with probability-of-success modeling, drug development decision analysis, or health economics
Comfort with LLMs or AI/ML pipelines in a production or research setting
Familiarity with dashboard/visualization tools (Streamlit, Plotly, Dash) and pipeline orchestration (Dagster, Airflow)
Healthcare OR finance domain knowledge is valued; both are not required.

Responsibilities

Work with the team to implement and maintain core portfolio engine: order management system, execution simulation layer, portfolio construction service, and performance tracking
Design risk frameworks that quantify exposure across a portfolio of drug development bets with radically different risk profiles, timelines, and failure modes
Run rigorous backtesting experiments with strict temporal constraints to evaluate Formation strategies against baseline approaches and measure marginal signal from new evidence sources
Coordinate across the organization to integrate internal Formation data sources (clinical trial data, genomic evidence, real-world data) and proprietary tooling into portfolio analytics pipelines
Work with product and engineering teams to build dashboards and reporting that communicate portfolio performance, risk metrics, and strategy comparisons to both technical and executive stakeholders
Collaborate with the broader data science team to ensure portfolio-level evaluation feeds back into model improvement and evidence prioritization