Senior Manager / Assistant Director - AI Researcher, …, Singapore

Role Overview This is a research role responsible for advancing
AI/ML-driven investment research
to support alpha generation and risk insight across the firm's quantitative and systematic investment platforms. This position is
research-led
and emphasizes: Novel signal discovery and validation
using modern machine learning. Strong experimental design, robustness and model risk discipline. Translating research into
investment-ready concepts , working with quant developers for productionisation. The AI Researcher will collaborate closely with portfolio managers, quant researchers, and risk partners to develop research that withstands
out-of-sample testing, regime shifts, and implementation frictions
(transaction costs, liquidity, capacity).
Key Responsibilities 1) Research Leadership in AI/ML for Investments Lead research initiatives applying AI/ML to alpha generation and risk insights across equities, fixed income, and/or multi-asset (depending on team mandate). Formulate hypotheses, design experiments, and drive research agendas focused on
signal stability, interpretability, and economic rationale . Evaluate and select modeling approaches (e.g., cutting edge deep learning algorithms, reinforcement learning) based on empirical evidence and implementation practicality. 2) Alpha Signal Discovery & Feature Research Create and test predictive features from structured and alternative datasets (e.g., prices, fundamentals, macro, curves/spreads, flows, options-implied measures, text/news). Research model families relevant to finance, including:
Time-series forecasting
and representation learning,
cross-sectional prediction
& ranking objectives, and
nonlinear factor discovery
and interactions. Develop frameworks for detecting and managing
non-stationarity
(regime shifts, concept drift, structural breaks). 3) Research Methodology, Robustness & Model Risk Discipline Establish and enforce rigorous research standards, including: leakage controls, realistic signal timing, corporate action adjustments walk-forward evaluation, time-series cross-validation, and stability diagnostics sensitivity testing across sub-periods, regimes, and market stress events. Diagnose and mitigate overfitting through sound regularization, feature selection discipline, and robust validation. Produce research documentation suitable for internal governance, including
assumptions, limitations, failure modes, and monitoring metrics . Contribute to model risk processes: validation support, explainability, and audit-ready research artifacts. 4) Portfolio & Implementation-Aware Research Translate model outputs into implementable signal definitions (ranking, scoring, forecasts) aligned to portfolio construction approaches. Incorporate practical constraints early: turnover, liquidity, transaction costs, latency/data availability, and capacity. Partner with portfolio construction and execution teams to ensure research remains robust after
cost and implementation adjustments . 5) Research Communication & Stakeholder Influence Present research findings to investment leadership with clarity: economic intuition, empirical results, and risk considerations. Contribute to the firm's though leadership by authoring and publishing sanitized AI research and methodological advancements in leading conferences and quantitative finance journals. Mentor and guide junior researchers on methodology, experimental design, and research hygiene. 6) Research-to-Production Collaboration Work with quant developers/engineering teams to transition validated research into production pipelines. Define requirements, acceptance criteria, and monitoring KPIs support post-launch research review and model drift investigations. Maintain an iterative research lifecycle: improvements, recalibration, and controlled retirement of decaying signals.
Required Qualifications, Skills & Capabilities Core AI/ML Research Skills (Required): Strong foundation in
statistical learning
theory. Expertise in time-series modelling and optimization. Practical experience with explainability and diagnostics (e.g., SHAP, permutation importance, stability tests) appropriate for investment oversight. Programming & Research Tooling: Advanced R or Python or Julia for research. Experience with deep learning frameworks (e.g. PyTorch / Flux.jl / TensorFlow). Strong research hygiene: Git, reproducible experiments, notebooks-to-library workflows, and structured documentation. Familiarity with experiment tracking tools (MLflow/W&B or equivalent) is beneficial. Data Competency: Strong skills in dataset curation, construction and labelling, including handling: survivorship bias, look-ahead bias, delayed data availability and corporate actions, missingness, outliers, vendor idiosyncrasies. Proficiency with
SQL
and working with large datasets comfort partnering with data engineering teams. Markets & Portfolio Context: Working understanding of market microstructure and implementation constraints (transaction costs, liquidity, slippage). Portfolio concepts: risk factors, diversification, drawdown, turnover, and capacity. Domain knowledge in at least one area (equities or fixed income) preferred.
Experience & Knowledge Required Education: Master's or PhD strongly preferred
in Machine Learning, Statistics, Computer Science, Applied Mathematics, Physics, Engineering, or related fields. Professional Experience: Typically,
6-8+ years
in ML/AI academic research, postdoc, quant research, or systematic investing (buy-side preferred strong sell-side or fintech acceptable). Demonstrated track record of
original research
that improved outcomes. Experience influencing research direction, mentoring others, and partnering with cross-functional stakeholders. Evidence of Research Depth: Peer-reviewed publications, patents, open-source contributions, or significant internal research outputs. Evidence of rigorous validation and an ability to explain why models work (or fail) across regimes. Key Competencies: Research leadership: sets direction, prioritizes high-impact questions, and drives rigor. Intellectual honesty and skepticism resists overfitting and 'backtest-first' thinking. Clear communication: simplifies complexity without overselling results. Collaboration: effective partner to PMs, risk, and engineering pragmatic about implementation realities.