VTAlgo LAB — HEFESTO · KRATOS · THEMIS · ZEUS Engine

2026 cohort

Five programs.

HEFESTO is the environment around the model. KRATOS is the control plane over the CLIs you already pay for. THEMIS is Financial Data Engineering. ZEUS Engine is the institutional research factory. AQUILES is Data Mining Rule Induction & Extraction Reloaded. They do not mix.

1 · ENVIRONMENT

HEFESTO

The system that controls the agents. Labs on the student’s repo.

→

2 · CONTROL PLANE

KRATOS

Seal the plan. Seat the table. No commit without a receipt.

→

3 · DATA

THEMIS

Financial Data Engineering & Point-in-Time.

→

4 · RESEARCH FACTORY

ZEUS Engine

One course. Six internal phases.

5 · RULES

AQUILES

Data Mining Rule Induction & Extraction Reloaded. Distinct program — not a ZEUS phase.

HEFESTO · Multi-Agent Harness Engineering

The system that controls the agents

10 phases · 15 modules · capstone 2 recorded tutorials / month ES audio · EN subtitles Student repo · not bridle

1 HARNESSModel → agent → harness

2 CONTEXTContext, state, tools

3 ROLESRole engineering

4 MULTI-AGENTOrchestration

5 VERIFYVerification + failure

6 MULTI-HARNESSAdapters, common governance

7 CONTROL PLANEDesign, not the Kratos product

8 GOVERNSecurity + observability

9 COSTTokens and equal-compute

10 CAPSTONEGoverned multi-harness system

Ten technical phases. Fifteen academic modules (I–XV) live inside them. Coursework clock is not sealed. Recorded tutorials: 2 / month (a different clock). Not Kratos.

KRATOS · Control plane

Not another agent

12 phases · 21 modules · capstone 2 recorded tutorials / month ES audio · EN subtitles Hands-on lab · Kratos / bridle

1 WIREKratos / bridle

2 CONTRACTSRequest, result, receipt

3 WORK UNITState machine

4 SEALLineup + human /seal

5 ADAPTERSNormalized I/O

6 PIPELINEImplementer → Reviewer

7 RECEIPTSContent-bound evidence

8 OPERATORDeterministic delivery

9 HEALTHQuotas and gates

10 SIDECARSResearcher / auditor

11 BOUNDSAgnosticism vs marketing

12 CAPSTONEGoverned work unit

Twelve technical phases. Twenty-one academic modules (I–XXI) live inside them. Coursework clock is not sealed. Recorded tutorials: 2 / month (a different clock). Do not reteach AGENTS.md here — that is HEFESTO.

THEMIS · Financial Data Engineering

Point-in-Time architecture.

6 phases · 12 modules · capstone Data contracts Not CHRONOS

I FOUNDATIONSFinancial Data Foundations

II PITTemporal & Point-in-Time Engineering

III SEMANTICSCanonical Market Semantics

IV REFERENCEFinancial Reference Infrastructure

V DOMAINSComplex Financial Domains

VI GOVERNANCEScientific Data Governance

Twelve academic modules · 36 chapters. Professional Capstone is the Financial Data Center Reference Implementation, outside 12/36. Coursework clock is not sealed.

ZEUS Engine · Institutional quant research

A factory. Not a trading system.

6 phases · 14 modules · capstone One professional course ES audio · EN subtitles

I ATENEAQuantitative Data Foundation

II APOLOQuantitative Data Runtime

III CHRONOSTemporal Research Governance

IV AEGISQuantitative Research Factory

V TALOSProduction Signal Runtime

VI ARTEMISAExecution Systems Engineering

One course. Phases are internal, not separate products. Orientation is outside 14/42. Coursework clock is not sealed. ARES is not taught here.

AQUILES · Data Mining Rule Induction & Extraction Reloaded

Induce explainable rules

5 modules + Validation layer AI tech 2026 Live cohort English

1 DATA ENGINEERINGPIT · gaps · quality

2 SAMPLING / BARSTime · volume · dollar

3 FEATURESPrice · flow · regime · path

4 SELECTIONIC · FDR · stability

5 RULE DISCOVERYInduction · RuleFit · ensembles

VALIDATION Transversal layer — not a sixth module. Data → Sampling → Features → Selection → Rules. leakage → temporal integrity → multiple testing → stability → OOS

Data Engineering

Intraday data. PIT / timestamps. Cleaning, gaps, multi-asset sync, quality.

Sampling / Bars

Time bars 1m–60m. Volume bars. Dollar bars. Compare sampling schemes.

Feature Engineering

Price/returns, volatility, volume/flow, structure, temporal, cross-asset, regime, path.

Feature Selection

Redundancy, dependence, IC, multicollinearity, FDR, temporal stability.

Rule Discovery

Induction, extraction, association, subgroup, sequential, trees → rules, RuleFit, ensembles.

Educational curriculum: 5 modules. Validation is a transversal layer, not a sixth module. Session counts and hours are not sealed.

Cohort 2026

Cinco programas.

HEFESTO es el entorno del modelo. KRATOS es el plano de control sobre los CLIs que ya pagan. THEMIS es ingeniería de datos financieros. ZEUS Engine es la fábrica institucional de research. AQUILES es Data Mining Rule Induction & Extraction Reloaded. No se mezclan.

1 · ENTORNO

HEFESTO

El sistema que controla a los agentes. Labs en el repo del alumno.

→

2 · PLANO DE CONTROL

KRATOS

Sellar el plan. Sentar la mesa. No hay commit sin receipt.

→

3 · DATOS

THEMIS

Ingeniería de datos financieros y Point-in-Time.

→

4 · FÁBRICA DE RESEARCH

ZEUS Engine

Un curso. Seis fases internas.

5 · REGLAS

AQUILES

Data Mining Rule Induction & Extraction Reloaded. Programa distinto — no es una fase de ZEUS.

HEFESTO · Multi-Agent Harness Engineering

El sistema que controla a los agentes

10 fases · 15 módulos · capstone 2 tutoriales grabados / mes Audio ES · subtítulos EN Repo del alumno · no bridle

1 HARNESSModelo → agente → harness

2 CONTEXTContexto, estado, tools

3 ROLESIngeniería de roles

4 MULTI-AGENTOrquestación

5 VERIFYVerificación + fallos

6 MULTI-HARNESSAdapters, gobernanza común

7 CONTROL PLANEDiseño, no el producto Kratos

8 GOVERNSeguridad + observabilidad

9 COSTTokens y equal-compute

10 CAPSTONESistema multi-harness gobernado

Diez fases técnicas. Quince módulos académicos (I–XV) viven dentro de ellas. Reloj de cursada no sellado. Tutoriales grabados: 2 / mes (otro reloj). No es Kratos.

KRATOS · Plano de control

No es otro agente

12 fases · 21 módulos · capstone 2 tutoriales grabados / mes Audio ES · subtítulos EN Lab práctico · Kratos / bridle

1 WIREKratos / bridle

2 CONTRACTSRequest, result, receipt

3 WORK UNITMáquina de estados

4 SEALLineup + /seal humano

5 ADAPTERSI/O normalizado

6 PIPELINEImplementer → Reviewer

7 RECEIPTSEvidencia atada al contenido

8 OPERATOREntrega determinista

9 HEALTHCuotas y gates

10 SIDECARSResearcher / auditor

11 BOUNDSAgnosticismo vs marketing

12 CAPSTONEWork unit gobernada

Doce fases técnicas. Veintiún módulos académicos (I–XXI) viven dentro de ellas. Reloj de cursada no sellado. Tutoriales grabados: 2 / mes (otro reloj). Acá no se reenseña AGENTS.md: eso es HEFESTO.

THEMIS · Ingeniería de datos financieros

Arquitectura Point-in-Time.

6 fases · 12 módulos · capstone Contratos de datos No es CHRONOS

I FOUNDATIONSFinancial Data Foundations

II PITTemporal & Point-in-Time Engineering

III SEMANTICSCanonical Market Semantics

IV REFERENCEFinancial Reference Infrastructure

V DOMAINSComplex Financial Domains

VI GOVERNANCEScientific Data Governance

Doce módulos académicos · 36 capítulos. El Professional Capstone es la Financial Data Center Reference Implementation, fuera de 12/36. Reloj de cursada no sellado.

ZEUS Engine · Research cuantitativo institucional

Una fábrica. No es un sistema de trading.

6 fases · 14 módulos · capstone Un curso profesional Audio ES · subtítulos EN

I ATENEAQuantitative Data Foundation

II APOLOQuantitative Data Runtime

III CHRONOSTemporal Research Governance

IV AEGISQuantitative Research Factory

V TALOSProduction Signal Runtime

VI ARTEMISAExecution Systems Engineering

Un curso. Las fases son internas, no productos aparte. Orientation queda fuera de 14/42. Reloj de cursada no sellado. ARES no se enseña acá.

AQUILES · Data Mining Rule Induction & Extraction Reloaded

Inducir reglas defendibles

5 módulos + capa Validation Reloaded Cohorte en vivo Inglés y español

1 INGENIERÍA DE DATOSPIT · gaps · calidad

2 MUESTREO / BARRASTiempo · volumen · dólar

3 FEATURESPrecio · flujo · régimen · path

4 SELECCIÓNIC · FDR · estabilidad

5 REGLASInducción · RuleFit · ensembles

VALIDATION Capa transversal — no es un sexto módulo. Data → Sampling → Features → Selection → Rules. leakage → integridad temporal → multiple testing → estabilidad → OOS

Ingeniería de datos

Datos intradía. PIT / timestamps. Limpieza, gaps, sincronización multi-activo, calidad.

Muestreo / Barras

Barras de tiempo 1m–60m. Barras de volumen. Barras de dólar. Comparación de esquemas de muestreo.

Ingeniería de features

Precio/retornos, volatilidad, volumen/flujo, estructura, temporal, cross-asset, régimen, path.

Selección de features

Redundancia, dependencia, IC, multicolinealidad, FDR, estabilidad temporal.

Descubrimiento de reglas

Inducción, extracción, asociación, subgroup, secuencial, árboles → reglas, RuleFit, ensembles.

Currículo educativo: 5 módulos. Validation es una capa transversal, no un sexto módulo. Sesiones y horas: no selladas.

Cohort map

Five programs.

HEFESTO

KRATOS

THEMIS

ZEUS Engine

AQUILES

The system that controls the agents

Not another agent

Point-in-Time architecture.

A factory. Not a trading system.

Induce explainable rules

Data Engineering

Sampling / Bars

Feature Engineering

Feature Selection

Rule Discovery

Cinco programas.

HEFESTO

KRATOS

THEMIS

ZEUS Engine

AQUILES

El sistema que controla a los agentes

No es otro agente

Arquitectura Point-in-Time.

Una fábrica. No es un sistema de trading.

Inducir reglas defendibles

Ingeniería de datos

Muestreo / Barras

Ingeniería de features

Selección de features

Descubrimiento de reglas