Arvind Tawker Microsoft Research Letter

I am writing to recommend Arvind Tawker for a position at Microsoft Research. Arvind has been a technician in my lab for eight months, but I have known him since he applied for graduate school at CMU in April. He had his heart set on CMU, and under normal circumstances, he would have been accepted into our program. Due to uncertainty in funding, CMU — and I believe many other schools — was overly conservative with the number of students admitted. We were lucky to have the opportunity to hire Arvind in our module assembly center as a post-bacc technician. We hired him with the understanding that he would apply to graduate schools again the following year. He has already been admitted into several top-tier schools, including CMU. I hope we can convince him to stay at CMU!

I am an experimental particle physicist working on the CMS experiment at the Large Hadron Collider. At CMU, we are constructing an upgrade to the CMS detector, the High-Granularity Calorimeter, which involves assembling 5,000 8-inch silicon modules – essentially radiation-hard digital cameras. Each module consists of a silicon sensor attached to a printed circuit board containing readout electronics and a base plate that provides cooling and structural stability. Module construction is performed using automated robotic systems to assemble components and make approximately 200 electrical connections per module. To facilitate this, we have established a small manufacturing center on campus capable of producing up to 24 modules per day, with production set to begin in early 2025.

Arvind has been critical in getting our lab ready to build modules. As one of four full-time team members, he has contributed to a wide range of projects, from implementing computer vision in module assembly to optimizing workflow for production using principles akin to industrial engineering. Arvind has been responsible for operating the main gantry robot used for module assembly; he has built modules and has developed software using pattern-recognition algorithms to automate the assembly process. This has required him to learn LabVIEW software and to modify programs originally developed by domain-expert engineers.

Arvind is our resident software and AI expert. He is the point person for designing, developing, and maintaining our computer vision algorithms, which lie at the heart of the automated assembly. He is also working to streamline our database management. This currently involves manual synchronization of data collected at CMU with a central repository at CERN. Streamlining this process is essential to achieving our target throughput rate.

In summary, Arvind is a strong candidate for this position. I will let the other writers speak to his proposed research program. I will just say he delivers what he promises; he is determined and has been a very productive member of my group.