Skip to content
Nahid Quader

ML engineer · Industrial AI · Interpretability

Hi, I’m Nahid.

I’m a machine learning engineer currently working on industrial AI, helping manufacturers reduce scrap rates using AI.

I’m also interested in understanding how neural networks learn: the concepts and mechanisms underneath. Especially how we can audit them for hidden or unsafe behavior, and in some cases intervene on specific mechanisms.

Research →PublicationsCV

Research interests

  • 01

    Mechanistic interpretability

    Reverse-engineering the circuits and features that transformers use to implement specific behaviours.

  • 02

    Superposition & sparse dictionaries

    How models pack more features than dimensions, and how sparse autoencoders can recover them.

  • 03

    Interpretability for safety

    Using model internals to detect, explain and steer behaviour in deployed systems.

All publications →

Applied Sciences (MDPI) · 2025

Reinforcement Learning Agent for Multi-Objective Online Process Parameter Optimization of Manufacturing Processes

Akshay Paranjape, Nahid Quader, Lars Uhlmann, Benjamin Berkels, Dominik Wolfschläger, Robert H. Schmitt, Thomas Bergs

Manufacturing lines have to balance several conflicting goals at once, such as reducing scrap while keeping the process stable, and the right machine settings have to be chosen in real time while production is running. We frame this as a multi-objective reinforcement learning problem: an agent trained with Multi-Objective Maximum a Posteriori Optimization (MO-MPO) proposes process parameters online.