Financial Services
Modernizing a High-Risk Derivatives Pricing Platform at Citi
De-risking critical pricing workflows where failure could mean revenue loss, regulatory exposure, or trading disruption.
Services Provided
High-Risk System Modernization, API-based Platform Modernization, Risk-aware Incremental Change
Product Type
Technologies Used
API-based Web Platform
Kotlin, React, Typescript
Product Type
API-based Web Platform
Technologies Used
Kotlin, React, Typescript
Project Highlights
Eliminated stale-model risk — traders no longer rely on manually downloaded scripts, removing a major source of pricing errors
Platform now supports 9 tools across 6 teams, with confidence that changes won't silently break downstream workflows
Zero disruption to live trading during the entire modernization
About
In derivatives trading, pricing systems sit directly on the path between market activity, revenue, and regulatory oversight. Small errors — stale models, inconsistent calculations, or unapproved changes — can have outsized consequences.
At Citi, pricing models were central to active trading workflows. Changing how those models were distributed or updated carried real risk: incorrect pricing, operational disruption, or compliance issues were not acceptable outcomes.
Def Method was engaged to modernize this pricing platform without introducing new failure modes, making necessary change possible again in a system where breaking things is expensive.
Challenge
Pricing models were distributed manually. Traders downloaded Python scripts to their local machines, creating a decentralized system where multiple versions of the same model could exist at the same time. This introduced several high-risk conditions:
Reliability risk — Traders could unknowingly run stale or inconsistent models, leading to incorrect pricing during live trading.
Revenue risk — In fast-moving markets, even small pricing discrepancies could translate directly into financial loss.
Compliance risk — Manual distribution made it difficult to ensure that all traders were using approved, auditable pricing logic.
The challenge was not simply that the workflow was inefficient. The deeper issue was that the system could not be safely changed. Any attempt to improve or scale the workflow risked disrupting active trading or introducing errors that would be costly to unwind. This is a common modernization failure mode: teams know change is required, but the system has grown too risky to touch.
Solution
We treated this as a high-risk modernization effort, not a feature build. Rather than rushing to replace the existing workflow, we focused first on removing the conditions that made change dangerous. Our approach centered on three principles:
De-risk before expanding — We eliminated the most dangerous failure modes — version drift, stale models, and manual distribution — before enabling broader usage.
Define safe boundaries for change — By centralizing pricing logic behind a controlled interface, we reduced the blast radius of future changes and made behavior predictable.
Preserve operational continuity — All changes were introduced incrementally, ensuring that active trading workflows were never interrupted.
We introduced a centralized, API-based pricing platform that became the single source of truth for pricing models. Pricing models are served centrally, eliminating local copies and version drift. Traders always access the latest approved model. Changes to pricing logic are controlled, auditable, and reversible. The system scales across tools and teams without introducing inconsistency.
Results
The modernization delivered immediate and long-term benefits. Traders no longer rely on manually downloaded scripts, removing a major source of pricing errors. The platform now supports 9 tools across 6 teams, with confidence that changes won't silently break downstream workflows. The system was modernized with zero disruption to live trading.
Most importantly, Citi now has a pricing platform that can continue to evolve without fear of outages, audit findings, or revenue loss — the core goal of modernization when breaking things is expensive.
Modernization fails when teams focus on building new systems before making change safe. In this case, success came from de-risking first — identifying where change was dangerous, removing those risks, and only then enabling broader modernization. This approach turned a fragile, high-risk workflow into a stable foundation for ongoing change.