Siddharth Agrawal
Machine Learning Engineer at Meta
Stony Stratford, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolSiddharth Agrawal is a machine learning engineer with 13 years of industry experience and over five years focused on applied ML and computer vision, primarily building production systems for the security domain at Calipsa and Motorola Solutions. He has end-to-end ownership of ML products—from data collection and active labelling to model prototyping, evaluation, and productionization—and led projects that reduced false alarms dramatically and cut server and training costs substantially. A strong software engineering background from Amazon and open-source contributions to TensorFlow-based probabilistic modeling (Edward) and the C++ mlpack library reflect his ability to bridge research and production code. He has driven rapid productization in startup settings (shipping a revenue-generating camera-monitoring product in eight months) and later scaled those capabilities inside a larger enterprise. Now at Meta, he continues to combine rigorous research with pragmatic engineering to deliver scalable, cost-efficient ML systems.
12 years of coding experience
9 years of employment as a software developer
M.E. Computer Science and Engineering, M.E. Computer Science and Engineering at Indian Institute of Science (IISc)BITS Pilani, Birla Institute of Technology and Science
Github Skills (14)
bayesian-methods10
machine-learning10
deeplearning-ai10
variational-inference10
probabilistic-programming10
deep-learning10
cpp10
tensorflow10
cplus10
python9
regression9
neural-network7
statistics6
data-science6
Programming languages (4)
Github contributions (5)
blei-lab/edward
Mar 2017 - Jan 2018
A probabilistic programming language in TensorFlow. Deep generative models, variational inference.
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 7 PRs, 16 comments in 10 months
Contributions summary:Siddharth contributed significantly to the `edward` repository, which focuses on probabilistic programming. Their work involved implementing and refining probabilistic matrix factorization examples. They also addressed several issues related to the core inference algorithms, including fixing calculations within the Metropolis-Hastings method and enabling regularization in various inference methods like KLqp, BiGAN, and WGAN. The user demonstrated skills in TensorFlow-based probabilistic modeling and variational inference techniques.
inferenceneural-networksmachine-learningprobabilistic-programmingdeep-generative-models
mlpack/mlpack
Jun 2014 - Aug 2014
mlpack: a fast, header-only C++ machine learning library
Role in this project:ML Engineer
ML EngineerContributions:15 commits, 1 comment in 1 month
Contributions summary:Siddharth primarily contributed to the implementation of a cosine tree, a crucial component for efficient nearest neighbor search within the machine learning library. Their work included the development of the cosine tree implementation and associated cosine node implementation. Furthermore, the user integrated QUIC-SVD and Reg SVD into the project. The user also added test cases to validate the functionality of the regularized SVD method.
regressionheaderdeep-learningscientific-computingc-plus-plus
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.