Agrima Bahl
Senior ML Engineer, Mentor, and Human Rights Data Analyst
agrimabahl
San Francisco Bay Area
Berkeley, California
Joined August 2025
Network
1.1K connectionsAKMTHA
DDTNTB
PPJLVS
NRMTSH
IKDPPS
CSEMAK
AFTCSZ
Summary
Agrima Bahl is a seasoned Machine Learning Engineer with a strong background in developing AI-driven solutions, currently contributing to Credit Karma's fintech innovations. Her expertise spans Generative AI, Deep Learning, and LLMs, and she is actively involved in mentoring aspiring professionals in these fields. adplist+2
Beyond commercial applications, Agrima has a unique commitment to data science for social good, demonstrated by her ongoing consultant role at the Human Rights Data Analysis Group (HRDAG). There, she applies her skills to critical record linkage and data analysis for human rights projects, such as the Colombia project. hrdag+1
Her early career involved significant contributions to high-performance computing at NERSC, Lawrence Berkeley National Laboratory, where she focused on optimizing Intel's Knight's Landing processor. This work highlights her foundational expertise in computational systems and performance analysis. theorg+2
Agrima is also a thoughtful writer, sharing her insights on machine learning, Python programming, and Git best practices through her Medium blog. Her posts cover topics from statistical error metrics to efficient code reproduction and Pandas deprecations. medium
Work
Education
Writing
MAPE v/s MAE% v/s RMSE: What does MAPE-puts-a-larger-penalty-on-negative-errors mean!?
November 1, 2019Elegant Python code reproduction of 'most common words from a story'
May 1, 2018A simpler, efficient, and elegant reproduction of a medium article to visualize most common words in a story.
Gitting to the next level in git
January 1, 2018Discusses advancing Git skills beyond basic commits, pushes, and checkouts.
Deprecation warnings in Pandas v 0.22 and how to fix them
November 1, 2017Explains how to address deprecation warnings in Pandas, emphasizing the importance of not ignoring them for future-proofing code.
Understanding Knight's Landing HIgh Bandwidth Memory Using the STREAM Benchmark
September 1, 2016A presentation on the performance of Intel's Knight's Landing processor and its high bandwidth memory.