Hiring Signal
Hiring Signal
Candidate profiles derived from observable work — not self-reported competencies. The platform tracks what candidates solve, how they implement it, and how they rank on difficulty-adjusted challenges. Employers inspect that work directly.
For Candidates
Your profile is built automatically from what you do on the platform. No CV writing. No keyword stuffing. Work speaks.
Verified skill areas
Mapped to platform taxonomy: Stochastic Calculus, Monte Carlo, C++ numerical programming, Calibration, etc.
Depth indicators
Average complexity of completed modules and challenges, weighted by peer review score.
Consistency signals
Contribution frequency, recency, and progression rate over time.
Peer standing
Reputation percentile within cohort. Based on review outcomes, not self-reported seniority.
Work product access
Employers see code, derivations, and challenge solutions directly — not candidate summaries.
Visibility is opt-in
Anonymous profiles by default. You decide when — and to whom — your profile is visible. Named profiles require explicit consent.
For Employers
Pre-screening built on observed skill, not claimed competencies. Searchable by verified skill area, depth level, language proficiency, and activity recency.
Target role types
Quant Developer — C++, Python, implementation depth, system design
Pricing Quant — SDEs, PDEs, MC, derivation fluency, edge cases
Risk Quant — Greeks, VaR/ES, sensitivity analysis, model risk
Quant Researcher — Statistical reasoning, ML, backtesting rigour
Structuring Quant — Exotics, term sheets, hedge construction
Algo / Market Making — Microstructure, execution, inventory models
How scores are derived
Topic ratings
Each candidate has a Glicko-1 rating per topic, starting at 1200. It updates after every quiz submission based on difficulty (Easy → 1000, Medium → 1200, Hard → 1400, Expert → 1600 opponent rating). Mastery percentage is derived directly from the rating, not from completion counts.
Challenge scoring
Code submissions are executed by a sandboxed judge (Judge0). A solution is accepted only when it produces correct output within CPU and memory limits. XP and rating updates apply on the first accepted solution — not on attempts.
Quiz accuracy
Per-topic quiz accuracy is tracked independently of rating. A session is considered passed at ≥70% correct. Both accuracy and Glicko rating are surfaced in the candidate profile.
Common questions
How do I know the challenges are interview-grade?
They were designed by someone who has conducted and sat quant desk interviews across tier-1 banks and hedge funds for 15 years. The calibration is not theoretical — it reflects what actually gets asked at the Pricing Quant, Quant Developer, and Risk Quant levels.
What stops candidates from gaming the system?
Challenge scores are based on machine-executed correctness, not self-reporting. Topic ratings are difficulty-weighted — easy problems move the rating less. XP is awarded once per problem, not per attempt. There are no self-described skill levels anywhere on the profile.
How reliable is the signal for a candidate with few submissions?
A thin profile is an honest thin profile. The Glicko model tracks rating deviation (RD) alongside rating — high RD signals low confidence in the estimate. Employers can see both, and filter by activity recency.
Is employer access live?
The scoring and profile infrastructure is live. The employer-facing search interface is in early build. Register interest to be notified when it opens.
Employer access is in early build. Candidate profiles are forming as the platform grows. Register interest to be notified when employer search opens.