Skip to content

Portfolio standard

A transcript is secondary. Completion means evidence another person can inspect.

# Wave Artifact Type Minimum evidence Reviewer Pass test
1 Wave 1 Formal specification + proof Proof/specification A nontrivial safety or correctness property with explicit assumptions and adversarial counterexample attempts. Formal reviewer No hidden assumptions survive challenge.
2 Wave 1 Calculus/vector model Derivation + numerical check One nonlinear physical/sensor system with local linearization, integral balance, units and error bounds. Math/physics reviewer Analytic and numerical predictions reconcile.
3 Wave 1 Linear algebra identifiability dossier Analysis + code Rank/nullspace/conditioning/SVD analysis of a sensor or inverse problem. Numerical reviewer Failure due to ill-conditioning predicted before computation.
4 Wave 1 Nonlinear dynamics study Simulation + oral defense State model, equilibria, phase portrait, stability, perturbations and model boundary. Dynamics/control reviewer Qualitative predictions precede simulation.
5 Wave 1 Probability/risk model Analysis + Monte Carlo Analytic and simulated risk with dependence/common-cause assumptions. Statistics reviewer Simulation and analytic result agree or discrepancy explained.
6 Wave 1 Causal target-trial redesign Causal design Turn an observational claim into estimand/DAG/target trial and sensitivity analysis. Causal inference reviewer Causal claim independent of estimator.
7 Wave 1 Numerical solver verification Scientific computing ODE/PDE/inverse solver with convergence, conditioning and residual/conservation checks. Numerical methods reviewer Separates model and numerical error.
8 Wave 1 Objective-function audit Optimization A real optimization where proxy gaming is demonstrated and objective/constraints redesigned. Optimization/policy reviewer Shows a metric improving while true goal worsens.
9 Wave 1 Signal/information system Signals/information Sampling/filter/channel or sensor-information design with noise and bandwidth limits. Signals reviewer Destroyed/recoverable information identified.
10 Wave 1 Closed-loop control system Control/estimation Unknown plant model, estimator, controller, robustness/failure injection. Controls reviewer Works under specified perturbations; failure boundaries stated.
11 Wave 1 Calibration/DOE package Metrology/experiment Raw data, uncertainty budget, factorial/controlled experiment and reproducibility instructions. Experimental reviewer Independent reproduction succeeds.
12 Wave 2 Electromechanical device model Physics/engineering Force/field/circuit/energy model tied to build/simulation. Physics/electrical reviewer Energy and force calculations close.
13 Wave 2 Thermodynamic hard-bound analysis Thermo Efficiency/entropy/transport/heat-rejection bound for a frontier system. Thermal reviewer No first/second-law or transport violation.
14 Wave 2 Material claim characterization plan Materials Mechanism, predicted signatures, alternative explanations, characterization ladder and reproducibility. Materials reviewer Extraordinary claim requires multiple independent observables.
15 Wave 2 Embedded instrument Hardware/software Sensor -> analog -> ADC -> real-time logic -> actuator + calibration/watchdog/fault log. Embedded reviewer Timing/noise/power budgets match behavior.
16 Wave 2 Mechanical/fluid design review Mechanical CAD or detailed design with loads, flow, fatigue, tolerances, failure and test plan. Mechanical reviewer Governing failure mode identified before optimization.
17 Wave 2 Biological systems model Bio systems Transport/metabolism/gene-network or tissue model with controls and measurement plan. Biology reviewer Mass/energy/information and biological context consistent.
18 Wave 2 Evolutionary robustness study Evolution/systems biology Stochastic/evolutionary escape or adaptation model. Evolution reviewer Design survives or explicitly fails under adaptation.
19 Wave 2 Human physiological control model Neuro/physiology BCI/artificial-organ/exoskeleton closed loop including adaptation and safety. Neuro/physiology reviewer User/physiology is part of loop, not an external nuisance.
20 Wave 2 10,000-unit manufacturing plan Manufacturing/operations Routing, takt/rate, WIP, bottleneck, yield, QA, downtime, maintenance, supply and unit cost. Manufacturing reviewer Throughput/cost/yield reconcile.
21 Wave 2 Integrated safety case Safety/reliability/security Hazard log, STPA/FMEA/reliability/threat model, monitor and shutdown criteria. Safety/security reviewer Critical claims trace to evidence or explicit assumption.
22 Wave 3 Frontier replication Research reproduction Independent reproduction of one current/open result in chosen track. Domain researcher Central claim survives or failure is clearly documented.
23 Wave 3 Mechanism-level extension Research extension Minimal change targeting a stable failure mechanism with preregistered metrics. Domain researcher Predicts where it helps and where it should not.
24 Wave 3 Integrated frontier system System closure Domain capstone carrying result through relevant physical/data/operational constraints. Cross-domain panel No hidden magical subsystem.
25 Wave 3 Frontier research paper Paper Abstract, methods, results, negative results, limitations, reproducibility, cost/compute/materials. External reviewer Claims match evidence.
26 Wave 4 Historical/source claim ledger Human-systems analysis Primary/secondary sources, competing account, changed conclusion. Historian/source critic Contrary evidence creates material revision.
27 Wave 4 Objective/power/institution audit Governance Objectives, stakeholders, decision rights, vetoes, recourse, distributional effects. Policy/institution reviewer Changes system design or deployment conditions.
28 Wave 4 Incident/drift reconstruction Organizational safety Escalation timeline, weak signals, authority, normalization and control structure. Safety/org reviewer Produces leading indicators/stop-work changes.
29 Wave 4 Wicked-problem portfolio Governance/decision Multiple formulations, incompatible values, reversible portfolio, monitoring and reformulation. Mixed panel Does not pretend one scalar objective resolves contested values.
30 Wave 4 Conflict/design-review record Communication Recorded/annotated disagreement, interests, evidence, decision and unresolved issues. Peer/team reviewer Clarity increases even if consensus does not.
31 Wave 4 Economics and real-options model Economics/capital Bottom-up cost, sensitivity, staged investment, abandonment/pivot option. Finance/ops reviewer Technical priority changes under economics when appropriate.
32 Wave 4 Compliance/standards matrix Regulation Primary authority, requirement, evidence, owner, verification and update trigger. Domain compliance reviewer Legal minimum distinguished from safety/ethics.
33 Wave 4 Reproducible public release Research integrity Clean-environment reproduction package, data provenance, limitations and negative-results appendix. Independent reproducer Central result reproduced without oral help.
34 Wave 4 Lab/company constitution Institution design Mission/non-goals, decision rights, review cadence, stop-work, conflict and evidence standards. Leadership/institution reviewer Protects dissent/evidence when founder/PI is wrong.
35 Final 136-week dossier + 12-month program Integration Selected best artifacts, thesis, strongest objections, uncertainty, research agenda, kill/pivot criteria. Three-person adversarial panel You can explain what not to build and why.