An AI interview coach for candidates breaking into investment banking. It runs realistic mock interviews, so candidates can practise as often as they need before the real thing.
Technical questions
M&A, DCF and LBO questions of the kind asked in first rounds and Superdays.
Behavioural and fit
Practice on the "why banking", "walk me through your CV" and team-fit questions, including HireVue-style recorded rounds.
Designed by bankers
The interviewers were designed by former bankers from bulge-bracket firms, so the questions and standards match what candidates will face.
Still growing
Built with a co-founder, 1learner continues to attract new users.
1analyst
- Status
- In development
- Coverage
- Listed companies
- Access
- Qualifying users, at launch
An AI research platform that thinks like a buy-side analyst. It reads company filings, earnings calls and research, compares them against each other to see what has changed, and builds a view on the business, its balance sheet and its valuation. For qualifying users, that view becomes a trade recommendation with the reasoning and the confidence level attached.
Comparative research
The most useful insight often comes from setting one document against another: this year's filing against last year's, or the earnings call against the numbers. 1analyst is built to find exactly that.
Numbers and judgements you can audit
Financial figures combine a deterministic calculation layer with LLM judgement. This focuses attention on the decisions that matter, and keeps both the numbers and the judgements behind them fully auditable.
Socratic due diligence
The engine carries out its own due diligence. A junior analyst agent answers the questions, and a senior analyst agent tests, challenges and pulls the answers together into a view.
Learns as it works
Recurring judgements become policies, so the same situation is handled the same way next time. Its knowledge base grows with every company it covers, including data buried in report charts.
A view on demand
Ask about a company and get an actionable view, even when some data is missing, with confidence stated honestly. The view refreshes as new information arrives.
Confidentiality built in
Confidential and price-sensitive information is flagged at the point it comes in, and never appears in output without an explicit decision to include it.
How people stay in control
The engine can move forward fully autonomously, with complete auditability and under human control. A human operator can change, override or modify any machine decision, and directs the broad policies and investment approach the engine follows.
1tradr
- Status
- Early trading, in development
- Markets
- Equities
- Style
- Systematic, event-driven and momentum
A systematic trading engine running event-driven and momentum strategies in equities. It turns machine-learning signals into disciplined trades and draws on 1analyst's research as one of its inputs, so trades rest on fundamentals as well as on price.
Early results
Early trading in live conditions, entirely out of sample, has been encouraging so far, with returns well ahead of the market benchmark. It is early days, and much more testing is needed.
Signals that respect costs
Models read market behaviour across several time frames and adjust to the prevailing regime. A trade is taken only when the expected edge is a clear multiple of what it costs to trade.
Every exit planned in advance
Each position has its exits defined before entry: a stop scaled to volatility, a profit target at key price levels, a reversal of the signal, or a time limit.
Tested before it is trusted
Strategies go through walk-forward testing and statistical significance checks, with strict safeguards against look-ahead bias, before they trade.
Explains itself
A live dashboard shows each decision as it happens: the main drivers behind the signal, a plain-language explanation, and how past predictions have fared.
A research loop that improves it
AI reviews the trade journal, proposes improvements, and a panel of independent AI judges assesses them. Changes are tested in isolation on unseen data before anything is promoted.
Operations-grade safety
An independent risk and control layer, durable trade records and reconciliation against the broker, built to keep working even if the dashboard or network fails.
How people stay in control
Trades can run fully automated, inside guardrails and risk limits that sit outside the machine's control, or be executed by a human hand.
Follow the work
Leave your email to hear about new projects, early access and white papers. Updates go out only when there is something worth reading.