What Happens When a Beginner Refuses to Stay a Beginner

2026-04-23 · Source: Machine Learning on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Novice, medium

Summary

The author recounts their journey building an AI/ML-powered CRM dashboard, starting with minimal knowledge of customer relationship management. They detail the process from data exploration, where they initially struggled with customer segmentation concepts like RFM, to feature engineering, which they found less daunting than anticipated. The project involved developing customer segmentation and lead scoring models, achieving an AUC of 0.66 for lead scoring. A significant part of the application's intelligence came from integrating Google Gemini for AI-powered follow-up recommendations and a custom chatbot, after encountering limitations with OpenAI's free API. The project culminated in deploying the application on Streamlit, overcoming challenges related to optimal cluster selection, customizing the chatbot interface with HTML/CSS, and managing Streamlit's session state.

Key takeaway

For data scientists and machine learning engineers embarking on their first end-to-end projects, embrace the iterative and messy nature of real-world development. Your practical experience in debugging and problem-solving, like managing Streamlit session states or customizing interfaces, will teach you more than any tutorial. Prioritize completing a functional application over striving for initial perfection, as "done" provides invaluable learning and a deployable asset.

Key insights

Practical project experience is crucial for understanding real-world machine learning challenges beyond structured course material.

Principles

• Experimentation drives learning.
• "Done" often surpasses "perfect."

Method

The project workflow involved data exploration, RFM analysis, feature engineering, model training for segmentation and lead scoring, integrating LLMs for recommendations, and deploying on Streamlit.

In practice

• Use RFM for customer segmentation.
• Customize Streamlit with HTML/CSS.
• Master Streamlit session state for complex apps.

Topics

• CRM Dashboard
• Customer Segmentation
• RFM Analysis
• Feature Engineering
• Lead Scoring

Best for: AI Student, Data Scientist, Machine Learning Engineer

Related on AIssential

• You're learning data science. You're not becoming one. — The gap between learning data science and becoming a practitioner is bridged by real-world "friction" that builds judgment.
• Day 30 of SPE Boot Camp on AI & Machine Learning: My 30-Day Learning Journey — Hands-on project-based learning significantly enhances practical AI/ML skill acquisition.
• Your First 90 Days as a Data Scientist — Successful data science onboarding prioritizes connections, domain context, and data knowledge to drive business impact.
• I trained a model. What is next? — Maximize ML project impact by systematically sharing code, models, and knowledge through public platforms and documentation.
• The Internship That Made Me Stop Practicing and Start Building — A structured data science internship bridges the gap between theoretical knowledge and practical, end-to-end project execution.
• The Biggest Mistake AI Beginners Make — Successful AI projects prioritize understanding a real-world problem before selecting models or algorithms.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning on Medium.