Selected Work
How Baemax approaches complex institutional problems.
A small selection of work, chosen to show how a problem is framed and how delivery is
disciplined — not to catalogue everything that has been built. Engagements are described at a level
that names no client, employer, vendor or counterparty, and no outcome is claimed that cannot be
substantiated.
Each entry carries one provenance label, so it is always clear whether you are
reading about a platform Baemax built, independent research, or delivered engagement work.
Delivered engagement
Multi-asset investment-management platform
For institutions replacing or consolidating the platform behind portfolio
management, trading and fund-administrator data.
The mandate
Baemax’s first engagement: to lead the design and implementation of a multi-asset
investment-management platform for a large institutional investor.
Approach
A specialist vendor platform combined with institution-specific data transformation, loading,
reconciliation, configuration, controls and operating processes.
Scope
From implementation design through fund-administrator data transformation, loading and
reconciliation, platform configuration, testing and operational readiness to controlled production
go-live.
Outcome
The platform went live to the agreed timetable and within the approved budget.
What is public
The institution, its data and the platform’s configuration are confidential and are not
described here.
Baemax platform
Execution governance — Baemax Aegis
For institutional teams responsible for trading workflows,
technology, operations and risk where order approval spans multiple desks and systems.
The problem
Between a trade decision and its execution, authority is often implicit. Approval lives in
messages, mail and convention, and the audit trail is reconstructed afterwards from systems that
were never designed to answer the question. When something goes wrong, the institution cannot
readily show who decided what, when, and on what basis.
The question
Who owns an order between intent and execution — and can that
ownership be evidenced afterwards without reconstruction?
Approach
Aegis treats governance as a design input rather than a reporting afterthought. It sits
between OMS, EMS and risk systems as a coordination layer rather than a replacement: orders carry
explicit approval state, decision ownership is role-based, and the audit trail is produced by the
workflow itself instead of assembled later.
Principles
- Sit between existing systems; replace none of them.
- Make authority explicit rather than conventional.
- Produce evidence as a by-product of the workflow, not a later reconstruction.
- Keep deterministic control over discretionary judgement, and record both.
What is public
An executive overview describes the platform’s governance model, operating posture and
current maturity.
Executive overview (PDF)
The application itself is not publicly accessible and is not linked from this site. To see
it in use, request a demonstration.
Independence
Baemax Advisory’s advice is vendor-independent. Recommendations are never conditional on using
Baemax software, and Aegis is not presented as deployed at any institution.
Research note
Designing systems that fail safely
For teams responsible for architecture, operational resilience and
trading risk.
The problem
Consolidation buys efficiency and quietly spends redundancy. Services that look
independent share a control plane, a region or a provider, so failures modelled as unrelated
arrive together. Resilience gets reported as uptime while the real exposure is correlation.
The question
How do we design trading systems that fail safely rather than
simply stop?
Approach
Begin from failure domains rather than availability statistics. Ask what remains true when
a dependency is gone, and design the degraded state deliberately — so that stopping safely is
intended behaviour with defined ownership, not an incident.
Principles
- Map failure domains, not uptime.
- Design the degraded mode explicitly; it is a mode, not an outage.
- Prefer recoverability over raw availability.
- Treat observability as a control, not a dashboard.
- Kill switches and hard stops are designed states with named owners.
Research
Baemax Lab holds the broader body of independent research, patent-pending inventions and working
papers. It is research shared for discussion rather than advisory work, which is why it sits outside
Selected Work.
Visit Baemax Lab