Experience

Research Engineer (August 2013 - Present):

I am a part of the Congitive Solutions and Services Group. The current focus of my work is to build analytical models integrated with Watson technology to generate business intelligence and insights that can improve the productivity and effectiveness of IT-Enabled services.

IBM Research - India, Bangalore

MRank: Ranking with Multiple Experts
Proposed a novel approach, MRank, which reduces the influence of unreliable and malicious experts in a multiple experts setting.
Provided theoretical analysis of our approach and bounds on the effect of evil experts on the learner and demonstrated the efficacy of the proposed approach through detailed experimental comparison with state-of-the-art algorithms on benchmark dataset
Selected in top 15 among all the interns’ projects at IRL
TRaNce: Trace Norm for Ranking
Developed a matrix factorization based framework, TRaNce, that employs a trace norm regularizer for learning to rank.
Provided a rigorous theoretical analysis encompassing generalization (both stability and consistency) bounds for the proposed methodology.
Also provided experimental evidence to corroborate the efficacy of our technique.

Worked on "High-Performance LINPACK on Graphics Processing Units"
- Ported High Performance Linpack on NVIDIA Tesla C1060.
- Proposed optimizations to achieve the desired speedup and the benefits of GPU.
- Division of tasks between CPU and GPU.
Poster accepted in Grace Hopper Celeberation (GHC) 2010

Vmukti Solutions Pvt. Ltd, Ahmedabad (incubated in IIM-A)

Worked on modules of two live projects, named "Vmukti Collaborative" and "Vmukti Call Center"