Podcast

Data Science Leadership: Product-First ML, Recommenders & RTB, MLOps, Hiring & Mentoring

S6E9

Open original DataTalks.Club episode

YouTube Spotify Apple Podcasts

data science machine learning MLOps leadership career growth team building hiring

Data Science Leadership: Product-First ML, Recommenders & RTB, MLOps, Hiring & Mentoring

Original Episode

Use these links for the canonical episode and media sources.

Episode Overview

How do you lead a data science team that prioritizes product impact while building recommender systems, real-time bidding (RTB) solutions, and maintainable MLOps? In this episode, Mariano Semelman, Head of Data Science at OLX Group with over 13 years of experience, walks through practical leadership decisions that bridge models and products.

People

Use these links to connect the episode to guest notes.

Mariano Semelman

Chapter Summary

Use these checkpoints to decide whether to open the source transcript.

0:00 - Podcast Introduction
1:26 - Guest Intro: Mariano Semelman, Head of Data Science at OLX
2:20 - Passion for Product Applications in Data Science
2:59 - Career Journey: Software Dev to Data Science Leadership
5:45 - Daily Responsibilities: Meetings, Mentoring & Planning
7:54 - Team Composition: Data Scientists and ML Engineers
8:39 - Transition to Management: First Team of Five
10:51 - Decision Mindset: Saying Yes and Learning on the Job
12:52 - 30-60-90 Plan: Onboarding, Listening, and Learning
15:16 - Project Onboarding: Rapid Learning and Trusting Reports
17:26 - Embracing Wrong Assumptions to Trigger Discussion
19:57 - Experience with Search and Recommender Systems
21:19 - Advertising Domain: Real-Time Bidding and Campaign Optimization
23:09 - Transferable DS Practices: Problem Framing & Feature Engineering
26:16 - Diagnosing Model Issues: Overfitting, Data, and Features
29:29 - Product-First Mindset: Prioritizing User Impact
30:06 - From Model to Product: Experiments, Deployment, Rules of ML
33:36 - Start Simple and Fail Fast: Iterative Testing in Production
36:12 - CRISP-DM Process and Deployment Realities
36:50 - Prioritization: Where Modeling Time Delivers Impact
40:25 - Feedback Timing: When to Give Performance Feedback
44:17 - Feedback Technique: Ask Permission, Care, and Offer Options
48:13 - One-on-Ones and Creating a Safe Growth Environment
50:39 - Handling Departures: Supporting Team Members Who Leave
52:37 - Technical Work as a Manager: Code Reviews and Prototypes
54:58 - Delegation and Architectural Involvement via Senior Engineers
55:48 - Hiring and Remediation: Interviews, Probation, Development Plans
1:01:37 - Staying Current: MLOps, NLP, and Engineering Best Practices
1:05:47 - Key Takeaways and Follow-Up Opportunities