UBS Warburg – Sesame Global Market Adaptor

My first role out of university placed me at the centre of one of the world’s most technically sophisticated trading environments. UBS Warburg’s Sesame platform was the core bus of the London trading floor — a Java and CORBA-based market adaptor that connected the bank’s internal trading systems to equity and derivative exchanges globally, routing order flow across every major market in the UK, Europe, US, and Japan.

Working on Sesame as a graduate trainee from 2000 to 2006 was the most formative experience of my career. The standards were exacting, the people were exceptional, and the problems were as real as it gets.

What Sesame Was

Sesame was the integration backbone of the UBS trading floor — a single, unified adaptor layer that standardised the interface between the bank’s proprietary trading tools and the full range of global exchange connectivity. Order flow from internal systems would pass through Sesame to reach exchange matching engines; market data and execution reports flowed back through the same channel.

The exchange coverage was global in scope:

• UK & Ireland — London Stock Exchange, LIFFE
• Europe — Euronext, CAC (Paris), Deutsche Börse (Frankfurt), Borsa Italiana (Milan), Bolsa de Madrid, and further Continental markets
• US — NYSE, NASDAQ, CME, and derivative venues
• Japan — Tokyo Stock Exchange and associated markets

For a trading floor operating simultaneously across all of these, Sesame was critical infrastructure. A degraded or unavailable adaptor meant degraded or unavailable trading — directly impacting the bank’s ability to execute on behalf of clients and on proprietary positions. Reliability and performance were not engineering goals; they were commercial imperatives.

My Contributions

I joined the Sesame team as a graduate trainee — the most junior person on a team of engineers working alongside quantitative analysts from Oxford, Cambridge, and the world’s leading universities. The calibre of that environment shaped how I think about software engineering permanently: rigorous, precise, performance-conscious, and utterly intolerant of ambiguity.

Auto-failover feature — my most significant technical contribution was helping design and implement the auto-failover mechanism that ensured continuous order routing under primary infrastructure failure. In a trading environment, the inability to route orders — even briefly — has immediate financial consequences. The failover system detected primary network failure and automatically rerouted order flow through an alternate path, maintaining trading continuity without manual intervention. Building this as a graduate, working within a system of that scale and consequence, was a formative lesson in what production reliability actually means.

Straight-through processing and automation — working with the proprietary trading API tooling exposed to the Sesame interface, I contributed to increasing the degree of straight-through processing across equity and derivative workflows — reducing manual touch points between order initiation and exchange execution. In a competitive market, every reduction in latency and manual intervention is a structural advantage, and the STP work was directly oriented towards that goal.

GL Trade integration — GL Trade was the order management system used by traders on the floor to manage positions and route orders. Working at the intersection of GL Trade and the Sesame API layer meant understanding both the trader’s operational requirements and the infrastructure constraints of exchange connectivity — a perspective that consistently proved valuable throughout the engagement.

3rd-line production support — providing 3rd-line support for the London trading desk meant owning the most complex production issues: those that couldn’t be resolved by the operations or 2nd-line teams. In a live trading environment, resolution speed is measured in minutes, not hours, and the quality bar for diagnosis and explanation is correspondingly high.

The Environment

The UBS trading floor in London in the early 2000s was operating at the frontier of what was technically possible in electronic trading. CORBA was the standard distributed computing middleware of the era — verbose and demanding by modern standards, but the right tool for the high-performance, typed, cross-language integration that financial systems required before the REST and messaging paradigms became dominant.

Working alongside quantitative analysts and traders at this level — people applying mathematics and computation to markets at the highest professional tier — set a standard for rigour that I’ve carried through every subsequent role. The expectation wasn’t just that the software worked, but that you understood precisely why it worked and could defend every design decision under scrutiny.

It was also where I first understood that latency isn’t a non-functional requirement — it’s a competitive weapon. The infrastructure investments and engineering discipline I’ve applied to Betfair exchange trading systems twenty years later trace directly back to what I learned on the Sesame team at UBS.

UBS FTSE / LSE Euronext