Skip to content

Instantly share code, notes, and snippets.

@Emmanuerl
Created December 13, 2025 02:19
Show Gist options
  • Select an option

  • Save Emmanuerl/523956cea52a1a0fbce941dd83e66cff to your computer and use it in GitHub Desktop.

Select an option

Save Emmanuerl/523956cea52a1a0fbce941dd83e66cff to your computer and use it in GitHub Desktop.
Engineering Intelligence: Keynote Talk at ML Lagos Community Day 2025

Engineering Intelligence

Building Practical ML Systems

Models don’t fail in notebooks.
They fail in production systems.

Why This Talk Exists

  • Most ML talks focus on:
    • Models
    • Metrics
    • Papers
  • Most ML failures come from:
    • Systems
    • Assumptions
    • Operations

This talk is about the part we keep ignoring.


The Reality Check

ML ≠ Intelligence
ML = A component in a system

  • A model is useless without:
    • Data pipelines
    • Serving infrastructure
    • Monitoring
    • Feedback loops

Intelligence only emerges when the system works.


A Simple Truth

Great models in bad systems fail.
Average models in great systems succeed.

  • Netflix
  • Google Search
  • Fraud detection systems

None rely on perfect models — they rely on well-engineered systems.


The ML Illusion

Common mistake:

“Once the model is good, we’re done.”

Reality:

  • Training is 10%
  • Everything else is 90%

From Notebook to Nightmare

Typical journey:

  1. Notebook works 🎉
  2. Model is deployed 🚀
  3. Traffic increases 📈
  4. Data changes 🔄
  5. Predictions degrade 📉
  6. Nobody notices 😬

This is not a modeling problem.
It’s a systems problem.


System Thinking for ML

Ask system-level questions:

  • Where does data come from?
  • Who owns data quality?
  • What happens when inputs change?
  • How do we know when the model is wrong?
  • What happens when the model is down?

ML as a Living System

ML systems are not static software.

They:

  • Age
  • Drift
  • Decay
  • Break silently

ML systems behave more like organisms than code.


Core Components of Practical ML

Every production ML system has:

  1. Data ingestion
  2. Feature generation
  3. Model inference
  4. Decision logic
  5. Monitoring & feedback

Ignore one → the system collapses.


Data Is the Real Model

Hard truth:

  • Your model learns your data bugs
  • Your model amplifies data bias
  • Your model reflects data staleness

Most “model issues” are actually data issues.


Training–Serving Skew

What you train on ≠ what you serve.

Causes:

  • Different pipelines
  • Different preprocessing
  • Missing values
  • Real-world messiness

Same code ≠ same data.


Prediction Is Not the Goal

Users don’t want predictions.

They want:

  • Decisions
  • Actions
  • Outcomes

Examples:

  • Fraud score ≠ fraud prevention
  • Recommendation ≠ conversion
  • Classification ≠ business value

Intelligence Needs Context

A model output without context is dangerous.

You need:

  • Confidence
  • Constraints
  • Business rules
  • Human overrides

Intelligence = Model + Rules + Judgment


Failure Modes You Must Design For

  • Bad inputs
  • Missing features
  • Cold start
  • Traffic spikes
  • Model timeouts
  • Partial outages

If you didn’t design for it, it will happen.


Observability for ML Systems

Traditional monitoring is not enough.

You must observe:

  • Input distributions
  • Output distributions
  • Decision rates
  • Business impact

Logs and metrics ≠ ML observability.


Silent Failure Is the Worst Failure

ML systems often:

  • Don’t crash
  • Don’t alert
  • Just get worse

A wrong prediction at scale is worse than downtime.


Feedback Loops Matter

Without feedback:

  • Models never improve
  • Errors repeat
  • Bias grows

Key question:

How does the system learn it was wrong?


Human-in-the-Loop Is Not a Weakness

Humans:

  • Catch edge cases
  • Provide corrections
  • Create trust

The best ML systems collaborate with humans, not replace them.


MLOps Is Systems Engineering

MLOps is not:

  • Just tools
  • Just pipelines
  • Just CI/CD

MLOps is:

  • Reliability
  • Ownership
  • Lifecycle management

Same principles as distributed systems.


Ownership Over Everything

Ask:

  • Who owns data quality?
  • Who owns model performance?
  • Who responds when it degrades?

If the answer is “the model”, you already lost.


Build Boring, Reliable Intelligence

Avoid:

  • Over-complex models
  • Fragile pipelines
  • Unnecessary automation

Prefer:

  • Simple models
  • Strong defaults
  • Clear fallbacks

Boring systems scale. Fancy ones break.


Practical Wins Over Academic Wins

In production:

  • Interpretability > +0.5% accuracy
  • Stability > novelty
  • Maintainability > brilliance

The ML Engineer Mindset Shift

Stop thinking:

“How good is my model?”

Start thinking:

“How resilient is my system?”


What Great ML Teams Do Differently

  • Design for failure
  • Monitor business impact
  • Ship small, iterate fast
  • Treat ML like infrastructure

Final Takeaway

Intelligence is engineered, not trained.

  • Models are components
  • Systems create value
  • Engineering makes ML real

Closing Quote

“In production, intelligence is not about how smart your model is —
it’s about how well your system survives reality.”

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment