Data Scientist Computer Vision

Sahil Sood

857-***-**** • *****.****@*******.*** • LinkedIn • Github

Summary

Data Scientist with 3+ years of experience delivering data driven solutions in telecom & retail through network, market, product analytics. Proficient in advanced machine learning methods including computer vision, data gathering, feature engineering, reporting tools and database ops. Skills

Tools: Python, YAML, SQL, R, GCP, Tableau, Alteryx, Airflow, Snowflake, Jira ML: TensorFlow, Keras, OpenCV, Scikit-learn, YOLO, ResNet, Autoencoders, Computer Vision, Hadoop, XGBoost Analytics: Causality, Cohort, Funnel, Trend, Experimentation, Segmentation, Clustering, PCA, Feature engineering Work Experience

Data Scientist Vritta, Delhi Jul 2023 - Dec 2023

Product recommendation & analytics for optimizing ad-spend, pricing and seasonal promotions.

• Developed computer vision-based recommendation model that detects and classifies fashion accessories from front-facing full-body images using K-nearest neighbors for similarity search.

• Implemented full-body image detection using YOLO, and feature generation by leveraging ResNet, transforming image data into high-dimensional embeddings to enhance classification accuracy.

• Improved RoAS by 15% by reorganizing ad budgets utilizing Marketing Mix Modeling (MMM) and evaluating advertising effectiveness through direct response and RoAS distribution curves.

• Led A/B testing initiatives for promotional campaigns, achieving a 9% boost in conversion rates and a 10% increase in average order value by pinpointing the most effective offers.

• Applied advanced pricing analytics techniques, enhancing revenue by 8% through price elasticity modeling and competitive benchmarking, resulting in a refined pricing strategy and a 12% uplift in LTV. Data Analyst Ericsson, Bengaluru Jun 2018 – Nov 2022 Customer-facing role, improving network operations with data engineering, technology consulting and RF teams.

• Designed a multi-class fault prediction framework with 76% field accuracy, integrating network counters and qualitative data, resulting in 15% reduction in post fault maintenance activities.

• Saved $300K in operational costs by optimizing tower crew deployment and reducing unnecessary site visits.

• Implemented a LSTM-AE based anomaly detector to predict X2 link degradation 24 hours in advance.

• Improved service efficiency in network operations by 30% by decomposing key processes, establishing KPIs, and implementing real-time dashboards for performance tracking and decision-making.

• Developed an 8-week customer service volume forecaster, reducing cost per contact by 15%. Projects

Ninja-cart Produce Tag April 2024 - Jun 2024

• Developed a computer vision based multi class image classification system using TensorFlow’s ResNet-50 implementation. Generalized the pre-trained model by implementing data augmentation, callbacks and dropouts, improving accuracy to 91% from 82%.

Data Analytics Capstone Intern Straumann Group April 2024 - Jun 2024

• Developed innovative product bundling strategies using association rule mining, identifying high-Lift product pairs, resulting in a 9% increase in field sales by optimizing cross-sell and upsell opportunities.

• Conducted in-depth sales performance analysis via cohort analysis and RFM segmentation, uncovering actionable insights that informed targeted marketing campaigns and enhanced conversion rates. Education

Northeastern University, Boston, MA Jan 2023 - Jun 2024 MS Analytics (Applied Machine Intelligence, Agile Product Management) JP University of Information & Technology Aug 2013 - Jun 2017 Bachelor of Technology (Electronics & Communication Engineering)

Contact this candidate