Job Summary:
Plutus21 is looking for a Quantitative Researcher to discover, test, and improve systematic alpha signals and portfolio construction for low-frequency equity strategies (typically daily to monthly horizons). This role is designed for exceptional quantitative thinkers coming from Physics, Mathematics, Statistics, Engineering, or other rigorous fields.
Location: Remote
Key Responsibilities:
Research and hypothesis generation:
- Translate investment ideas into testable hypotheses with clear metrics and failure criteria
- Build simple baselines first, then iterate toward stronger models only when justified
- Data and features (research-grade)
- Work with panel/time-series equity data and build features with strict as-of availability
- Implement careful data checks (missingness, outliers, corporate actions, calendar alignment)
Evaluation and robustness:
- Design validation protocols appropriate for time series (walk-forward, rolling windows, cross-sectional splits)
- Detect and prevent common research pitfalls: look-ahead bias, leakage, overfitting, multiple comparisons
- Perform robustness analysis: turnover, drawdowns, concentration, regime sensitivity, stability across time and cohorts
- Backtesting and portfolio construction
- Implement or extend low-frequency backtests for signals and portfolios
- Model basic frictions realistically (transaction costs, slippage assumptions, liquidity/turnover constraints)
- Collaborate with engineering/trading to productionize the strongest research findings
Communication and collaboration
- Write clear research memos: what you tried, what worked, what didnt, what you recommend next
- Present results transparently, including uncertainty, limitations, and risk considerations
Qualifications (Core):
- Strong quantitative foundation in probability/statistics and at least one of: linear algebra, optimization, numerical methods
- Ability to design experiments and reason about measurement (baselines, controls, uncertainty, sanity checks)
- Ability to write working analysis code in Python (preferred) or another language, and communicate code/results clearly
- Comfort with real-world messy datasets and non-stationary behavior
- Strong written communication and intellectual honesty (you can say this is inconclusive and explain why)
- Prior research experience (academic, industry, independent) demonstrating end-to-end ownership
- Evidence of strong software fundamentals even without formal CS training: readable code, modularity, reproducibility
- Work involving time-series or observational data where leakage is a risk (forecasting, causal inference, experiments)
Nice to Have (Not Required):
- Any exposure to markets, equities, factor models, or portfolio construction (we can teach this)
- Familiarity with common research tools: numpy/pandas/scipy/statsmodels/sklearn, Jupyter, Git
- Experience with simulation/Monte Carlo, Bayesian methods, or causal inference