Nikhil Thakur profile photo 1 on Tal

Nikhil Thakur

Software Engineering @ Capital One, Cisco | University of Cambridge 2021, Indian Institute of Technology (BHU) Varanasi. 2018

India

8+ years of work experience

Capital One
Jan 20269 mosPresent

Bengaluru

Cisco
May 20259 mosJan 2026

Bengaluru, Karnataka, India

Revamping media services for FedRamp and sovereign cloud
Cisco
Oct 20231 yr 8 mosMay 2025

London, England, United Kingdom

Media Engineering - Webex Backend. Developing the transcoded experience for both Webex and Cloud Video MS Teams Interop (CVI) solutions of Cisco.
Cisco
Dec 20211 yr 11 mosOct 2023

London, England, United Kingdom

Media Engineering UK - Webex Media Backend
University of Cambridge
Oct 20201 yr 3 mosDec 2021

Cambridgeshire, England, United Kingdom

This research work involved the development of a Parallel Multigrid Algorithm used for solving PDEs to remove both the Low and High-frequency components of errors by using multi-level grids hierarchy and linear grid transfer operators. This algorithm was developed in SYCL / DPC++ and was tested for a grid having 177mn DoFs. The parallel multigrid implementation was tested on a multicore Intel Xeon CPU, and a parallel speedup of 26 was obtained on a 28 core system, indicating a parallel efficiency of 86 %. The results confirmed the successful development of the prototype of parallel multigrid using SYCL. After further improvements (kernel fusion and consistent memory management), a fully parallel implementation of this algorithm will be ready. It will be integrated with open source FEniCS / DOLFINx to check the performance against state-of-the-art solvers already present for commercial use.
ISRO - Indian Space Research Organization (U.R. Rao Satellite Center)
Jul 20182 yrs 2 mosAug 2020

Bengaluru, Karnataka, India

• Worked on the development of electro-optic payloads and laser-driven sensors (altimeters, rangers, velocity meters, etc.) for various satellites, rovers, and other spacecraft launched for defense and communication purposes. • Responsible for the design and computational analysis of these payloads (frequency response, random response, etc.), end-to-end fabrication, quality checks at design and fabrication stages, assembly, and field tests to ensure their performance against set ISRO benchmarks. • Tested several sensors and devices used in the Chandrayan - 2 moon landing mission of ISRO • Worked on the development of indigenous atomic clocks, handled the design and analysis phase. • Acting as a member of the SPADEX Design Review Committee (SDRC) within ISRO, analyzed various mission intricacies w.r.t. various optical sensors and mechanical payloads to be used in the mission. • Guided 2 research interns in the research works of Laser Range Finder and Laser Altimeter used in SPADEX and Moon-Landing Mission.
Technical University Munich
May 20173 mosJul 2017

Munich, Bavaria, Germany

• Worked on implementing Convoluted Neural Network (CNN) in TensorFlow to predict an accurate difference method and calculate numerical derivatives at a given point for a data set. This model was, in turn, used for calculating the derivative of a flow streamline, used for velocity prediction in a turbulent flow field. • Generated the training and testing dataset for the CNN model in Python • Performed a study for tuning the feature filters for an in-depth understanding of the CNN model to improve the prediction accuracy and visualized its accuracy and error over a span of 1000 epoch on the TensorBoard platform.

Nikhil’s AI Stack

antigravity
antigravity
claude code
claude code
gemini
gemini
hugging face
hugging face
windsurf
windsurf

Education

University of Cambridge

MPhil in Engineering, Computational Mathematics and Parallel Algorithms

University of Cambridge

20202 yrs2022
Indian Institute of Technology (Banaras Hindu University), Varanasi

Bachelor of Technology - BTech, Mechanical Engineering

Indian Institute of Technology (Banaras Hindu University), Varanasi

20144 yrs2018

Join Nikhil and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Nikhil Thakur on Tal

Headline

Software Engineering @ Capital One, Cisco | University of Cambridge 2021, Indian Institute of Technology (BHU) Varanasi. 2018

About

Experienced scientist with a demonstrated history of working in the deep-space research industry. Skilled in C++, Python, HPC, and Machine Learning. Strong research background with a Bachelor of Technology - B.Tech focused in Mechanical Engineering (focussed on Computational Science) from the Indian Institute of Technology (Banaras Hindu University), Varanasi, and a Master of Philosophy in Engineering from the University of Cambridge in High-Performance Computing. As a side hustle, I mostly find myself struggling with specific areas of Machine Learning, stock investment, Chess, reading all kinds of stuff, and exploring new trends.

Location

India

AI stack

antigravity

claude code

gemini

hugging face

windsurf

Work experience

Software Engineering at Capital One, Jan 2026 - Present · 9 mos, Bengaluru

Software Engineer at Cisco, May 2025 - Jan 2026 · 9 mos, Bengaluru, Karnataka, India, Revamping media services for FedRamp and sovereign cloud

Software Engineer at Cisco, Oct 2023 - May 2025 · 1 yr 8 mos, London, England, United Kingdom, Media Engineering - Webex Backend. Developing the transcoded experience for both Webex and Cloud Video MS Teams Interop (CVI) solutions of Cisco.

Software Engineer at Cisco, Dec 2021 - Oct 2023 · 1 yr 11 mos, London, England, United Kingdom, Media Engineering UK - Webex Media Backend

Post Graduate Researcher at University of Cambridge, Oct 2020 - Dec 2021 · 1 yr 3 mos, Cambridgeshire, England, United Kingdom, This research work involved the development of a Parallel Multigrid Algorithm used for solving PDEs to remove both the Low and High-frequency components of errors by using multi-level grids hierarchy and linear grid transfer operators. This algorithm was developed in SYCL / DPC++ and was tested for a grid having 177mn DoFs. The parallel multigrid implementation was tested on a multicore Intel Xeon CPU, and a parallel speedup of 26 was obtained on a 28 core system, indicating a parallel efficiency of 86 %. The results confirmed the successful development of the prototype of parallel multigrid using SYCL. After further improvements (kernel fusion and consistent memory management), a fully parallel implementation of this algorithm will be ready. It will be integrated with open source FEniCS / DOLFINx to check the performance against state-of-the-art solvers already present for commercial use.

Scientist/Engineer - 'SC' at ISRO - Indian Space Research Organization (U.R. Rao Satellite Center), Jul 2018 - Aug 2020 · 2 yrs 2 mos, Bengaluru, Karnataka, India, • Worked on the development of electro-optic payloads and laser-driven sensors (altimeters, rangers, velocity meters, etc.) for various satellites, rovers, and other spacecraft launched for defense and communication purposes. • Responsible for the design and computational analysis of these payloads (frequency response, random response, etc.), end-to-end fabrication, quality checks at design and fabrication stages, assembly, and field tests to ensure their performance against set ISRO benchmarks. • Tested several sensors and devices used in the Chandrayan - 2 moon landing mission of ISRO • Worked on the development of indigenous atomic clocks, handled the design and analysis phase. • Acting as a member of the SPADEX Design Review Committee (SDRC) within ISRO, analyzed various mission intricacies w.r.t. various optical sensors and mechanical payloads to be used in the mission. • Guided 2 research interns in the research works of Laser Range Finder and Laser Altimeter used in SPADEX and Moon-Landing Mission.

Research Intern at Technical University Munich, May 2017 - Jul 2017 · 3 mos, Munich, Bavaria, Germany, • Worked on implementing Convoluted Neural Network (CNN) in TensorFlow to predict an accurate difference method and calculate numerical derivatives at a given point for a data set. This model was, in turn, used for calculating the derivative of a flow streamline, used for velocity prediction in a turbulent flow field. • Generated the training and testing dataset for the CNN model in Python • Performed a study for tuning the feature filters for an in-depth understanding of the CNN model to improve the prediction accuracy and visualized its accuracy and error over a span of 1000 epoch on the TensorBoard platform.

Education

MPhil in Engineering, Computational Mathematics and Parallel Algorithms, University of Cambridge

Bachelor of Technology - BTech, Mechanical Engineering, Indian Institute of Technology (Banaras Hindu University), Varanasi