Applied Agentic Limited was founded by Terence Cheng, a former Portfolio Manager at CPP Investments with 14+ years of buy-side systematic equity experience. He brings the auditable research process, factor risk discipline, and regulated-environment governance of institutional capital markets to the firm’s agentic AI infrastructure.
Terence leads delivery full-time — scoped pilots through production deployment, with the same practitioner who designs the workflow operating the system alongside your team.
The company sits at the intersection of capital markets fluency and applied AI engineering — building the auditable, governed, human-in-the-loop infrastructure that regulated finance and professional services firms need to deploy LLMs in production.
Earlier career includes R&D at Algorithmics, cryptography software at Research In Motion (BlackBerry), and large-scale systems work at Sun Microsystems.
Systematic equities — areas of expertise
Cash equities strategies
- Long/short market-neutral, cross-sectional, factor investing; long-only index-enhanced
Single-name option strategies
- Vol capture, theta harvesting
Equity research
- Single-name option factors, variance swaps, implied vol skew
Engineering
- Alternative data, feature engineering, quant model building
- Axioma / scenario generation / portfolio optimization
- End-to-end production systems
Portfolio management
- Proprietary + Barra risk models
- Black-Litterman; portfolio attribution
- Weekly rebalance; daily model
Trading
- Execution research
- OMS/EMS implementation
- Algo development
- TCA — implementation shortfall, market impact, slippage
- Almgren-Chriss, VWAP, POV, TWAP
Education & credentials
- Master of Finance (MFin) — Queen’s University
- Master of Management Sciences (MMSci) — University of Waterloo, Faculty of Engineering
- Bachelor of Mathematics (BMath), double major in Computer Science and Combinatorics and Optimization — University of Waterloo
Agent platform
Sylia — agentic orchestration
Sylia is the agentic orchestration layer: it supervises and synthesizes outputs across LLMs in a distributed multi-agent network, coordinating models, tools, and human-in-the-loop checkpoints in the case studies on this site.
Built for complex, multi-step workflows — voice-led client intake, RAG-backed retrieval, presentation synthesis, and governed pipeline triggers. A product system for regulated teams, not a human hire or substitute for licensed professional advice.