Akash Dhaka
AI Scientist Benchmarking Engineer at Silo AI
Espoo, Finland
Summary
🤩
Rockstar🎓
Top SchoolAkash Dhaka is an AI scientist and benchmarking engineer with 11 years of experience, currently optimizing inference serving and NPU model compilation at AMD/Silo AI in Espoo. He holds a PhD in Probabilistic Machine Learning from Aalto University with first-author publications in top venues like NeurIPS and was a NeurIPS 2023 top reviewer, reflecting continued community service. His work bridges research and production: from Gaussian processes, Bayesian optimization and variational inference to practical LLM finetuning, quantization (GPTQ) and ONNX/DAG compilation for faster inference. He also contributed to the well-known SheffieldML/GPy repository by implementing new likelihoods and EP tests, evidencing deep probabilistic modeling expertise alongside systems engineering. Notably, he pairs low-level hardware-aware optimization with a strong mathematical foundation, enabling measurable gains in real-world model throughput and uncertainty-aware ML applications.
11 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Machine Learning, Statistics, Doctor of Philosophy (Ph.D.), Machine Learning, Statistics at Aalto UniversityMaster's Degree, Machine Learning, Master's Degree, Machine Learning at KTH Royal Institute of TechnologyIndian Institute of Technology RoorkeeMaster's degree, Mathematics and Statistics, Master's degree, Mathematics and Statistics at University of Helsinki
English, Hindi, Swedish
Github Skills (7)
machine-learning10
probability-distribution10
python10
distributions10
gaussian-processes10
testing9
scipy9
Programming languages (4)
Github contributions (5)
SheffieldML/GPy
Jun 2017 - Aug 2017
Gaussian processes framework in python
Role in this project:
ML Engineer
ML EngineerContributions:22 commits, 3 PRs, 8 comments in 2 months
Contributions summary:Akash primarily contributed to the GPy framework by implementing and testing new likelihood functions, specifically focusing on the LogLogistic and Weibull likelihoods. Their work involved creating and modifying code for these functions, including calculations for probability density functions, gradients, and Hessians. They also added and refined test cases, particularly for Expectation Propagation (EP) methods, demonstrating a focus on the application of Gaussian processes and related inference techniques.
gaussiangaussian-processespython
adhaka/adhaka.github.io
Dec 2017 - Sep 2020
Contributions:27 commits, 16 pushes, 1 branch in 2 years 9 months
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.