About Me

GSAK are the initials of my name, Guruvayur Sasikumar Aakash Krishna.

I am a Masters student at the University of Alberta. My interests are in applications of AI in real world systems. I am currently working under the supervision of Dr. Omid Ardakanian on applying Reinforcement Learning for building control.

View my CV

Work Experience

Graduate Research Assistant, University of Alberta

May 2022 — Present

Helped develop a novel RL-based algorithm for reducing cost of training building Heating Ventilation and Air Conditioning (HVAC) system control policies
Our algorithm reduces energy consumption by up to 30\% depending on the building
Used Weights and Biases to track and visualize experiments
Published our work in peer-reviewed journals and conferences

Research Engineer, TCS Research and Innovation

Jul 2019 — Aug 2021

Built an Electric Vehicle (EV) user behavior and traffic simulator for the city of Luxembourg
Developed dynamic pricing strategies to manage EV charging demand and supply
Developed a "Supply Uberization" solution, and published our work in ACM e-Energy `20
Developed and implemented Reinforcement Learning (RL) algorithms for dynamic pricing
Published the work on dynamic pricing with RL to IEEE Intelligent Transportation System Magazine

Research Intern, TCS Research and Innovation

Jan 2019 — Jun 2019

Worked on Cyber Physical Modelling for solar PhotoVoltaic (PV) cells
Built a test suite for for analyzing behavior of large scale solar PVs, e.g., solar farms
This test suite helped generate data to validate existing, and develop new fault detection and classification algorithms

Intern, Thermo Fisher Scientific

May 2018 — Jul 2018

Responsible for building an automation solution for the Thermo Scientific Spinnaker robotic arm
Built a custom dataset and trained a CNN based object detector to help it identify specific locations on different Thermo Scientific instruments
This robot removes the need for saving these locations manually and automates the entire process for transporting hazardous chemicals from/to various biomedical instruments
Optimized the deep learning model to run on a low-power embedded system

Education

Masters in Computing Science (Thesis)

University of Alberta

2021 — present

Wrote a thesis titled "Policy Selection for Transfer Learning in the Building Control Domain"
Worked with Dr. Omid Ardakanian and Dr. Matthew E. Taylor
Specialization in applied reinforcement learning (RL)
Recipient of the Alberta Graduate Excellence Scholarship (AGES)