Pengyi Interview Team
PIT / Company workspace

Interview practice,
organized by company.

PCT supplies current company and role diligence. PCV supplies the selected, evidence-bounded CV. PIT turns both into timed simulations, fresh problems, scorecards and repair work.

Directories
08
Active packages
11
Runnable mocks
23
Measured mocks
00

Eight preparation directories

Interview labs

Company- and role-family preparation / fictional cases / explicit evidence gates

001

Jane Street interview lab

Run all eight Package 004 role labs: TDOE, Infrastructure Automation, Data Engineer, Quantitative Research, Strategy & Product, MLE, MLR and SWE.

Designed 8 role branches 8 runnable mocks
002

General Quant role-family lab

Package 006 tests research validity, Python implementation, portfolio construction, risk, and a final evidence defense without pretending one firm has one universal process.

Designed Quant role family 1 runnable mock
003

HKUDS research application lab

Run the compact PhD 001 research-fit simulation. Package 007 adds the complete proposal, experiment and supervisor-panel defense.

Designed Research application 2 runnable mocks
004

Tencent Hunyuan interview lab

Package 008 separates Agent Harness, RL Framework, and RAG/Agent engineering into three evidence-bounded simulations; current role activity still requires a manual official-page check.

Role check required 3 engineering tracks 3 runnable mocks
005

Anzi Fund interview lab

Package 010 separates Quant IT, Trading Algorithm Research and High-Frequency Quant Data into three Chinese-first, evidence-bounded simulations.

Vacancy check required 3 quant tracks 3 runnable mocks
006

Unicode Digital interview lab

Package 011 tests Factor Research, Quant Development and Low-Latency Trading Systems against official role requirements and explicit eligibility gates.

Eligibility gates active 3 quant systems tracks 3 runnable mocks
007

RadixArk AI infrastructure lab

Package 013 separates the AI Infra Resident primary route from the Developer Experience stretch route, with cache, serving-benchmark, cookbook and evidence gates.

Official roles reviewed 2 AI infra tracks 2 runnable mocks
008

Recursive automated research lab

Package 014 tests experiment lineage, repeated-seed decisions, evaluator integrity and a falsifiable engineering expression of interest without inventing a vacancy.

No named public role Research systems 1 runnable mock

Operating protocol

Do not memorize. Expose the gap.

  1. 01Select one role and one CV.
  2. 02Solve a fresh problem out loud.
  3. 03Score evidence, judgment and correction.
  4. 04Build one bounded repair artifact.