Open roles
one mandate each · no overlapping ownership
R-01
{{ statusLabel }}
Quantitative Researcher
Find structure at the resolution it exists — and carry it to production yourself.
ResearchSingapore · HK
R-02
{{ statusLabel }}
Quant Developer — Systems Architecture
Own the trading platform end to end. Infrastructure is part of the intelligence.
EngineeringSingapore · HK
R-03
{{ statusLabel }}
Portfolio Manager
Take a mandate, a risk budget and a capital line. Own the outcome.
TradingSingapore · HK
R-01 · RESEARCH
Quantitative Researcher
A prediction is not a result. You will own an idea from raw market state through to capital at risk, and you will still own it when it decays.
TeamResearch
LevelMid to senior
Reports toHead of Research
What you will own
01
Choose the resolution a market should be observed at, and defend the choice — price, queue, flow, timing, venue mechanics, or the relationships between them.
02
Build the full path: data integrity, state representation, model, cost and impact assumptions, validation, live deployment.
03
Reconcile predicted against realized — markout, slippage, capture — and act on the gap rather than explaining it.
04
Detect decay early and retire strategies without sentiment.
05
Convert a finding into shared capability — a dataset, a validation gate, a reusable feature — so the next researcher starts further along.
What we look for
Strong probability, statistics and time-series foundations, and the judgment to know when they do not apply.
Fluent Python and enough production discipline that your research code can survive contact with live capital.
Real familiarity with microstructure — order books, queue dynamics, fee and rebate mechanics, latency, adverse selection.
Evidence that you have shipped something into live markets and were accountable for how it behaved.
Matters less than people assume
A finance degree, a named-firm pedigree, or a publication record. We read the reasoning, not the letterhead. Backgrounds in physics, applied mathematics, computer science and engineering are all common here.
R-02 · ENGINEERING
Quant Developer — Systems Architecture
The platform is not a service layer under the research. It decides what the research is able to see, how fast it can act, and how quickly a bad state is caught.
TeamPlatform & execution
LevelSenior to staff
Reports toHead of Systems
What you will own
01
The architecture from venue connectivity through market data, order management, risk gates and post-trade reconciliation.
02
Latency where it changes the outcome, and correctness everywhere — including the paths that only run when something has already gone wrong.
03
Observability as a first-class product: time to detect, time to diagnose, time to recover, each measured and driven down.
04
The research-to-production path — simulation fidelity, deployment gates, kill switches, safe rollback.
05
Adding venues and instruments without adding a proportional amount of operational burden.
What we look for
Systems-level C++ or Rust, plus Python — and strong opinions about where each belongs.
Concurrency, memory behaviour, networking and measurement done properly — you profile before you claim.
Experience running production systems that lose money when they are wrong, and the operational instincts that follow.
Architectural taste: the judgment to keep a system that a small team can still reason about in a year.
Matters less than people assume
Prior trading-industry experience. Strong engineers from adjacent real-time domains adapt quickly; we would rather teach markets than teach rigour.
R-03 · TRADING
Portfolio Manager
You receive a mandate, a risk budget and a capital line, and you carry the result. Automation scales your reach; it does not dilute your accountability.
TeamTrading
LevelEstablished track record
Reports toRisk & Capital Committee
What you will own
01
A defined strategy mandate: which markets, which mechanisms, which risks you are paid to take.
02
Sizing, exposure and drawdown inside an explicit risk budget — with the limits visible to you before they bind.
03
The full P&L attribution of your book — alpha, execution cost, financing, fees — not the headline number.
04
A working partnership with research and platform: you specify what the book needs, and you are accountable for using it.
05
Scaling into new markets as capacity allows — and saying so plainly when capacity is reached.
What we look for
A verifiable systematic track record, with risk and capacity described as precisely as return.
A clear account of where the edge comes from, and an honest view of the conditions that end it.
Enough technical depth to work directly with researchers and engineers rather than through a translator.
Discipline under drawdown — the temperament to cut, wait, and re-enter on evidence.
Matters less than people assume
The absolute size of the book you last ran. We care how the return was produced, how it behaved when conditions turned, and whether you can describe both without rounding.
How we hire
four stages · two to three weeks · one decision maker per stage
Read01
We read your work, not your formatting. A short note on something you built beats a cover letter.
Depth02
Ninety minutes on one problem you know deeply — yours, not ours. We push until we find the edge of it.
Real work03
A problem from our actual queue, on our data, with our constraints. Paid, and scheduled around your life.
Decide04
A decision and a reason, either way, within five working days. No silent pipelines.
Send the work, not the résumé.
One page on a problem you owned end to end — what you believed, what the data said, what you shipped, and what it did once it was live.
We hire on evidence and judgment alone. If a stage of our process would be difficult for you for any reason, tell us and we will adapt it.
DIMENSION BLADE
DIMENSION REVEALS STRUCTURE · BLADE TURNS JUDGMENT INTO ACTION
Dimension Blade is a proprietary trading firm that trades solely for its own account. It does not manage external capital or provide investment advice, brokerage services, or investment products to the public.
© 2026 DIMENSION BLADE