Get Started for Free →
Learn Programs Library About Us Contact Get Started for Free →
Requires: Applied Machine Learning + Python Fundamentals
In beta · Coming soon

Machine Learning & GenAI
for Finance.

Build a three-layer cash flow forecast on a real pharma case study. Legacy baseline, static ML, dynamic walk-forward retraining, and a guardrailed GenAI loop. This is what production-grade finance ML looks like.

9 modules
~5 hours hands-on
64 sub-videos
Real Excel artifact
GitHub Codespaces
Lifetime access
$249at launch
Not open to the public yet. Join the waitlist for first access.
vertex_pharma_cash_synthetic.csv19 cols · 247 rows
weekcash_actual_mtrial_velocityshock_flag
1124.210.831
1133.870.711
1144.020.791
Three-layer forecast comparison
$3.0M$3.5M$4.0M$4.5Mw100w110w120w130COVID-19 shock · w112
ActualLegacy baselineStatic MLDynamic ML
MAE comparison
Legacy Baseline0.847← walk-forward
Static ML0.612← walk-forward
Dynamic ML0.391← walk-forward
Improvement53.8%
The case

A real forecast. A real shock. A real Excel model.

Case briefingSynthetic · Week 100-130
RoleFP&A Analyst, mid-cap pharma
CompanyVertex Pharmaceuticals (synthetic data)
ChallengeWeekly cash forecast that holds up during disruption
Current toolBlended seasonal model in Excel
ProblemThe Excel model breaks during the COVID-19 shock (Week 112)
Your jobBuild something better. Three layers of improvement.
01Layer 1

Legacy Baseline

Translate the Excel model into Python. Match every formula. Prove the audit trail.

02Layer 2

Static Machine Learning

Fit a gradient boosting model on operational drivers. Beat the baseline. Watch it drift.

03Layer 3

Dynamic Walk-Forward

Retrain as new data arrives. Recover from the shock. This is the showpiece.

+ GenAI feature engineering loop
Outcomes

From Excel artifact to production forecast.

01
Excel
C112:
=OFFSET(B112,-52,0)
Python
# baseline.py
cash.shift(52)

Translate Excel to Python

Reproduce the legacy Excel baseline in Python, cell by cell. Watch the numbers match. Build the audit trail.

02
walk-forward loop
predictobserveappendretrain

Build a walk-forward ML loop

Code the walk-forward retraining loop from scratch. Predict, observe, retrain. This is how production ML teams maintain models.

03
scenarios · cfo summary
scenariocash $MMAE
Base3.420.391
Downside2.870.612
Upside3.940.422

Run scenarios for the CFO

Bump procurement delays and promo spend. Generate base, downside, and upside scenarios. Output a PDF the CFO can read.

04
guardrails.py
PromptSpec(
allowed_columns=["cash", "trial_*"],
max_features=3,
)
assert schema.validate(features)

Use GenAI with guardrails

Use AI to suggest new features, but constrain it. Pydantic schemas, column allowlists, validation asserts. Finance-grade governance, not vibes.

Curriculum

9 modules. 64 sub-videos. ~5 hours of hands-on work.

📹 = recorded video⚙ = student work time
·01Walk-forward loop, from scratch
·02Conditional retraining triggers
·03Leakage guards
·04Retrain cadence as hyperparameter
·05Three-way forecast comparison
Prerequisites

This product is built for AML graduates.

This workshop assumes you have completed PyFi Applied Machine Learning or have equivalent experience with:

  • ·01scikit-learn pipelines
  • ·02train_test_split
  • ·03Mean absolute error
  • ·04Gradient boosting
  • ·05Hyperparameter tuning

If any of these are unfamiliar, complete Applied Machine Learning first. This is not a suggestion. The ML modules (M3 onward) will not make sense without this foundation.

Your tools

No install. No IT department. Open your browser and code.

GitHub Codespaces

Fully configured, browser-based, 2 minutes to launch. No installs, no IT tickets.

Real Excel artifact

Vertex_Baseline_Model.xlsx — the blended seasonal model you will translate, formula by formula.

Synthetic dataset

vertex_pharma_cash_synthetic.csv — 19 columns of weekly cash, operational drivers, and shock flags.

Pre-built src/ package

Reporting, charting, and PDF plumbing already written. You focus on the ML, not the boilerplate.

github.dev/pyfi/vertex-workshop
Explorer
▾ workshop
▸ data/
▸ src/
▾ notebooks/
M2.ipynb
M5.ipynb
M7.ipynb
Cell [4] · Walk-forward loop
for i in range(n_steps):
pred = model.predict(X[i])
history.append(obs[i])
if i % 4 == 0:
model.fit(history)
✓ Cell run · 0.84s · MAE 0.391
In beta · Coming soon

Be first in line.

Machine Learning & GenAI for Finance is in private beta while we finish it. Join the waitlist for first access and founding-member pricing before it opens at $249.

Machine Learning & GenAI for Finance

4.9out of 5
$249.00 USDat launch
Join the waitlist

No charge today. We'll email you the moment the next cohort opens.

What's included
9 modules, 64 sub-videos
~5 hours hands-on work
Real pharma case study with Excel artifact
Walk-forward ML (the showpiece)
GenAI feature engineering with guardrails
GitHub Codespaces environment
Lifetime access
30-day money-back guarantee
Questions

Frequently asked questions.

Yes. The ML modules assume familiarity with scikit-learn pipelines, train/test splits, MAE, and gradient boosting. If those are new, take AML first.

Get Started for Free →
Learn Programs Library About Us Contact Get Started for Free →
Requires: Applied Machine Learning + Python Fundamentals
In beta · Coming soon

Machine Learning & GenAI
for Finance.

Build a three-layer cash flow forecast on a real pharma case study. Legacy baseline, static ML, dynamic walk-forward retraining, and a guardrailed GenAI loop. This is what production-grade finance ML looks like.

9 modules
~5 hours hands-on
64 sub-videos
Real Excel artifact
GitHub Codespaces
Lifetime access
$249at launch
Not open to the public yet. Join the waitlist for first access.
vertex_pharma_cash_synthetic.csv19 cols · 247 rows
weekcash_actual_mtrial_velocityshock_flag
1124.210.831
1133.870.711
1144.020.791
Three-layer forecast comparison
$3.0M$3.5M$4.0M$4.5Mw100w110w120w130COVID-19 shock · w112
ActualLegacy baselineStatic MLDynamic ML
MAE comparison
Legacy Baseline0.847← walk-forward
Static ML0.612← walk-forward
Dynamic ML0.391← walk-forward
Improvement53.8%
The case

A real forecast. A real shock. A real Excel model.

Case briefingSynthetic · Week 100-130
RoleFP&A Analyst, mid-cap pharma
CompanyVertex Pharmaceuticals (synthetic data)
ChallengeWeekly cash forecast that holds up during disruption
Current toolBlended seasonal model in Excel
ProblemThe Excel model breaks during the COVID-19 shock (Week 112)
Your jobBuild something better. Three layers of improvement.
01Layer 1

Legacy Baseline

Translate the Excel model into Python. Match every formula. Prove the audit trail.

02Layer 2

Static Machine Learning

Fit a gradient boosting model on operational drivers. Beat the baseline. Watch it drift.

03Layer 3

Dynamic Walk-Forward

Retrain as new data arrives. Recover from the shock. This is the showpiece.

+ GenAI feature engineering loop
Outcomes

From Excel artifact to production forecast.

01
Excel
C112:
=OFFSET(B112,-52,0)
Python
# baseline.py
cash.shift(52)

Translate Excel to Python

Reproduce the legacy Excel baseline in Python, cell by cell. Watch the numbers match. Build the audit trail.

02
walk-forward loop
predictobserveappendretrain

Build a walk-forward ML loop

Code the walk-forward retraining loop from scratch. Predict, observe, retrain. This is how production ML teams maintain models.

03
scenarios · cfo summary
scenariocash $MMAE
Base3.420.391
Downside2.870.612
Upside3.940.422

Run scenarios for the CFO

Bump procurement delays and promo spend. Generate base, downside, and upside scenarios. Output a PDF the CFO can read.

04
guardrails.py
PromptSpec(
allowed_columns=["cash", "trial_*"],
max_features=3,
)
assert schema.validate(features)

Use GenAI with guardrails

Use AI to suggest new features, but constrain it. Pydantic schemas, column allowlists, validation asserts. Finance-grade governance, not vibes.

Curriculum

9 modules. 64 sub-videos. ~5 hours of hands-on work.

📹 = recorded video⚙ = student work time
·01Walk-forward loop, from scratch
·02Conditional retraining triggers
·03Leakage guards
·04Retrain cadence as hyperparameter
·05Three-way forecast comparison
Prerequisites

This product is built for AML graduates.

This workshop assumes you have completed PyFi Applied Machine Learning or have equivalent experience with:

  • ·01scikit-learn pipelines
  • ·02train_test_split
  • ·03Mean absolute error
  • ·04Gradient boosting
  • ·05Hyperparameter tuning

If any of these are unfamiliar, complete Applied Machine Learning first. This is not a suggestion. The ML modules (M3 onward) will not make sense without this foundation.

Your tools

No install. No IT department. Open your browser and code.

GitHub Codespaces

Fully configured, browser-based, 2 minutes to launch. No installs, no IT tickets.

Real Excel artifact

Vertex_Baseline_Model.xlsx — the blended seasonal model you will translate, formula by formula.

Synthetic dataset

vertex_pharma_cash_synthetic.csv — 19 columns of weekly cash, operational drivers, and shock flags.

Pre-built src/ package

Reporting, charting, and PDF plumbing already written. You focus on the ML, not the boilerplate.

github.dev/pyfi/vertex-workshop
Explorer
▾ workshop
▸ data/
▸ src/
▾ notebooks/
M2.ipynb
M5.ipynb
M7.ipynb
Cell [4] · Walk-forward loop
for i in range(n_steps):
pred = model.predict(X[i])
history.append(obs[i])
if i % 4 == 0:
model.fit(history)
✓ Cell run · 0.84s · MAE 0.391
In beta · Coming soon

Be first in line.

Machine Learning & GenAI for Finance is in private beta while we finish it. Join the waitlist for first access and founding-member pricing before it opens at $249.

Machine Learning & GenAI for Finance

4.9out of 5
$249.00 USDat launch
Join the waitlist

No charge today. We'll email you the moment the next cohort opens.

What's included
9 modules, 64 sub-videos
~5 hours hands-on work
Real pharma case study with Excel artifact
Walk-forward ML (the showpiece)
GenAI feature engineering with guardrails
GitHub Codespaces environment
Lifetime access
30-day money-back guarantee
Questions

Frequently asked questions.

Yes. The ML modules assume familiarity with scikit-learn pipelines, train/test splits, MAE, and gradient boosting. If those are new, take AML first.

Get Started for Free →
Learn Programs Library About Us Contact Get Started for Free →
Requires: Applied Machine Learning + Python Fundamentals
In beta · Coming soon

Machine Learning & GenAI
for Finance.

Build a three-layer cash flow forecast on a real pharma case study. Legacy baseline, static ML, dynamic walk-forward retraining, and a guardrailed GenAI loop. This is what production-grade finance ML looks like.

9 modules
~5 hours hands-on
64 sub-videos
Real Excel artifact
GitHub Codespaces
Lifetime access
$249at launch
Not open to the public yet. Join the waitlist for first access.
vertex_pharma_cash_synthetic.csv19 cols · 247 rows
weekcash_actual_mtrial_velocityshock_flag
1124.210.831
1133.870.711
1144.020.791
Three-layer forecast comparison
$3.0M$3.5M$4.0M$4.5Mw100w110w120w130COVID-19 shock · w112
ActualLegacy baselineStatic MLDynamic ML
MAE comparison
Legacy Baseline0.847← walk-forward
Static ML0.612← walk-forward
Dynamic ML0.391← walk-forward
Improvement53.8%
The case

A real forecast. A real shock. A real Excel model.

Case briefingSynthetic · Week 100-130
RoleFP&A Analyst, mid-cap pharma
CompanyVertex Pharmaceuticals (synthetic data)
ChallengeWeekly cash forecast that holds up during disruption
Current toolBlended seasonal model in Excel
ProblemThe Excel model breaks during the COVID-19 shock (Week 112)
Your jobBuild something better. Three layers of improvement.
01Layer 1

Legacy Baseline

Translate the Excel model into Python. Match every formula. Prove the audit trail.

02Layer 2

Static Machine Learning

Fit a gradient boosting model on operational drivers. Beat the baseline. Watch it drift.

03Layer 3

Dynamic Walk-Forward

Retrain as new data arrives. Recover from the shock. This is the showpiece.

+ GenAI feature engineering loop
Outcomes

From Excel artifact to production forecast.

01
Excel
C112:
=OFFSET(B112,-52,0)
Python
# baseline.py
cash.shift(52)

Translate Excel to Python

Reproduce the legacy Excel baseline in Python, cell by cell. Watch the numbers match. Build the audit trail.

02
walk-forward loop
predictobserveappendretrain

Build a walk-forward ML loop

Code the walk-forward retraining loop from scratch. Predict, observe, retrain. This is how production ML teams maintain models.

03
scenarios · cfo summary
scenariocash $MMAE
Base3.420.391
Downside2.870.612
Upside3.940.422

Run scenarios for the CFO

Bump procurement delays and promo spend. Generate base, downside, and upside scenarios. Output a PDF the CFO can read.

04
guardrails.py
PromptSpec(
allowed_columns=["cash", "trial_*"],
max_features=3,
)
assert schema.validate(features)

Use GenAI with guardrails

Use AI to suggest new features, but constrain it. Pydantic schemas, column allowlists, validation asserts. Finance-grade governance, not vibes.

Curriculum

9 modules. 64 sub-videos. ~5 hours of hands-on work.

📹 = recorded video⚙ = student work time
·01Walk-forward loop, from scratch
·02Conditional retraining triggers
·03Leakage guards
·04Retrain cadence as hyperparameter
·05Three-way forecast comparison
Prerequisites

This product is built for AML graduates.

This workshop assumes you have completed PyFi Applied Machine Learning or have equivalent experience with:

  • ·01scikit-learn pipelines
  • ·02train_test_split
  • ·03Mean absolute error
  • ·04Gradient boosting
  • ·05Hyperparameter tuning

If any of these are unfamiliar, complete Applied Machine Learning first. This is not a suggestion. The ML modules (M3 onward) will not make sense without this foundation.

Your tools

No install. No IT department. Open your browser and code.

GitHub Codespaces

Fully configured, browser-based, 2 minutes to launch. No installs, no IT tickets.

Real Excel artifact

Vertex_Baseline_Model.xlsx — the blended seasonal model you will translate, formula by formula.

Synthetic dataset

vertex_pharma_cash_synthetic.csv — 19 columns of weekly cash, operational drivers, and shock flags.

Pre-built src/ package

Reporting, charting, and PDF plumbing already written. You focus on the ML, not the boilerplate.

github.dev/pyfi/vertex-workshop
Explorer
▾ workshop
▸ data/
▸ src/
▾ notebooks/
M2.ipynb
M5.ipynb
M7.ipynb
Cell [4] · Walk-forward loop
for i in range(n_steps):
pred = model.predict(X[i])
history.append(obs[i])
if i % 4 == 0:
model.fit(history)
✓ Cell run · 0.84s · MAE 0.391
In beta · Coming soon

Be first in line.

Machine Learning & GenAI for Finance is in private beta while we finish it. Join the waitlist for first access and founding-member pricing before it opens at $249.

Machine Learning & GenAI for Finance

4.9out of 5
$249.00 USDat launch
Join the waitlist

No charge today. We'll email you the moment the next cohort opens.

What's included
9 modules, 64 sub-videos
~5 hours hands-on work
Real pharma case study with Excel artifact
Walk-forward ML (the showpiece)
GenAI feature engineering with guardrails
GitHub Codespaces environment
Lifetime access
30-day money-back guarantee
Questions

Frequently asked questions.

Yes. The ML modules assume familiarity with scikit-learn pipelines, train/test splits, MAE, and gradient boosting. If those are new, take AML first.