Pavan Kumar Unnam Email & Phone Number

Cupertino, California, Us

San Francisco, California, Us

San Francisco, California, Us

New York, Ny, Us

- Built a C library for l1 trend filtering technique.- Worked on trend filtering for time series data using l1 regularization as a function of time.

Pittsburgh, Pennsylvania, Us

- Built a topic segmentation model for detecting the topic change in speech transcripts.- Implemented a sequence classification model using Fasttext word embeddings.- Developed a model for topic summarization from transcripts.

New York, Ny, Us

-Teaching Assistant under Prof. Xiaofu He for graduate level course "Applied Machine learning for Image Analysis"-Grade assignments for a class of 44 students

New York, Ny, Us

-Teaching Assistant under Prof. Adam Kelleher (Cheif Data Scientist for Research at Barclays) for graduate level course "Machine learning products for Data Science"-Graded assignments for a class of 34-Conducted weekly office hours to assist students in udnerstanding the course material

Chennai, Tamilnadu, In

- Did research on Decision Support Systems using data from aviation industry- Applied advanced statistical and data science skills to conduct exploration and research on the data

Gurgaon, Haryana, In

Ad Relevance Engine [PostgreSQL, Python, Django]- Developed a relevance engine in Django for real estate advertisements using elastic search server- The engine uses user intents (from preference modeling) and a modified collaborative filtering technique which we named as "Intent Matrix Factorization"; Website conversion rate was raised by 30% and bounce rate was reduced by 40%Similar Entities on Housing website [MongoDB, Python, Django]- Built Web App in Django to serve similar flats using Mongo database; Application uses K-means to cluster similar flats- Intra cluster similarity score is calculated from normalized Euclidean distance between the flat features- Increased the popularity of the website thus increasing property listings by dealers Monitoring Relevance of Search Results [R, Shiny, Elastic Search]- Developed a dashboard in R to analyze the efficacy of the search relevance algorithm and to visualize user engagement - Devised sampling strategies to match demand and supply; Proposed use of Bandit algorithms for algorithm testing- Enabled property developers to offer best bids for a given search rank resulting in increased profits (~10 million INR) Auto-suggest while typing (like google search) [MongoDB, Python, Django]- Built an app in Django to aggregate past user search behavior and compute relevance score to sort search suggestions- Integrated product promotions with search result sorting which increased marketing revenues of the company

Charlotte, North Carolina, Us

Manufacturing Data Analytics - Developed a statistical model in R which analyzes the influence of input features of the Fiber Reinforced Plastic molding process on the output (strength and stiffness) using Design Of Experiments modeling techniques Data mapping - Mapped huge amount of data between two dissimilar meshes ​ 10 million data points - Developed a code in Matlab to map and for tensor transformation which saved around INR 5 lakh​ for​ the company from buying mapping tools

Master'S Degree, Data Science

Bachelor’S Degree, Mechanical Engineering

High School, Science