Shreyas Mahimkar

Senior Data Scientist / Engineer

• 1. Developed market mix modeling (MMM) framework using Bayesian combiners to synthesize national lift metrics with geo-specific ad performance data (e.g., Boston, Chicago), improving marketing spend attribution accuracy by 40%. This methodology directly quantified incremental ROI across channels, enabling a $1M revenue deal by demonstrating measurable campaign impact to enterprise clients 2. Operationalized first Reach Frequency Model using SparkSQL for US-scale projections, enabling digital clean room activation and driving $1M+ client contract 3. Designed and deployed a linear attribution system integrating causal impact analysis with multi-touch attribution modeling, achieving adoption by 10+ enterprise clients within the first quarter of product deployment 4. Won company hackathons for batch campaign reporting system using A/B testing with Intent-to-Treat (ITT) and Per Protocol (PP) causal frameworks, later productized to onboard 6+ enterprise clients through standardized experiment analysis pipelines 5. Established best practices for experiment QA through SparkSQL optimizations that reduced processing time by 60% across all client implementations1

Data Scientist / Enginner

• 1. Built and implemented may data pipelines for multiple clients, resulting in significant improvements in impression, reach, lift and attribution metrics. In this fast-paced startup environment, I leveraged my skills in data engineering, data science, product management, DevOps, and UI to wear multiple hats and ensure project success. 2. Developed Bash shell scripts to automate system maintenance, data processing workflows, and batch job scheduling using tools like cron and tmux reducing manual errors by 40%

Data Scientist / Enginner

• 1. Leveraged Random Forest algorithms to predict user-level program/channel viewership probabilities, achieving 30% accuracy improvement over previous models 2. Designed predictive analytics solutions using Linear Regression to forecast aggregated TV ratings, informing programming and advertising strategies 3. Conducted statistical correlation analysis on 10M+ viewer records to identify 15+ key behavioral patterns and demographic drivers of TV engagement 4. Established data-driven insights framework guiding $2M+ annual content acquisition decisions through feature importance analysis 5. Built hybrid TV advertising data models across SQL (MySQL), NoSQL (Presto), and cloud storage (S3), reducing query latency by 40% through indexing optimization 6. Architected star schema for ad performance tracking, enabling cross-platform campaign analysis across 50+ metrics 7. Engineered Scala Spark ETL pipelines processing 1.2TB daily of Nielsen ratings data, achieving 99.98% uptime SLA 8. Implemented S3-based data lake with Athena SQL interface supporting 150+ concurrent analysts through partitioned Parquet storage Data Modelling: Created TV Ad data model in NOSQL and SQL databases in Presto, MYSQL and S3. Created TV ratings data model in Scala Spark and storing it into Amazon S3 and accessing it via Amazon Athena. Data Analysis: Extracted actionable insights on TV viewership data (Big data) using zeppelin and matplotlib

Masters Assistant

• Candidate for Masters of Science in Computer Science May 2016 Related Courses: Parallel Data Processing in Map-Reduce, Machine Learning, Algorithms Projects: • Fighting Crime with Big Data – Java, Amazon Web Services – Elastic Map Reduce, S3, Weka: Developed a tool that predicts the crime location in Boston and Chicago streets by running multiple Map-Reduce programs on large dataset of crime using K-D tree Algorithm.

Data Science Intern

Senior Software Engineer

• Worked for a Financial Corporation – Auto Loans for a client in US 1. Designed and optimized PL/SQL packages, procedures, and functions to streamline data retrieval processes, improving query performance by 35%. 2. Developed ETL workflows using PL/SQL for data transformation and migration between Oracle databases, ensuring compliance with business logic for report generation. 3. Resolved critical defects in an Auto-Loan System by debugging PL/SQL scripts and Java code, reducing production incidents by 50%. 4. Automated data movement tasks using Python scripts, integrating with Oracle databases to reduce manual processing time by 25%. 5. Collaborated with cross-functional Agile teams to deliver new system features, including Unix shell scripts for job scheduling and batch processing. 6. Conducted database performance tuning by analyzing execution plans, indexing strategies, and materialized views, cutting latency by 20%. Worked for a Financial Corporation – Auto Loans for a client in US 1. Designed and optimized PL/SQL packages, procedures, and functions to streamline data retrieval processes, improving query performance by 35%. 2. Developed ETL workflows using PL/SQL for data transformation and migration between Oracle databases, ensuring compliance with business logic for report generation. 3. Resolved critical defects in an Auto-Loan System by debugging PL/SQL scripts and Java code, reducing production incidents by 50%. 4. Automated data movement tasks using Python scripts, integrating with Oracle databases to reduce manual processing time by 25%. 5. Collaborated with cross-functional Agile teams to deliver new system features, including Unix shell scripts for job scheduling and batch processing. 6. Conducted database performance tuning by analyzing execution plans, indexing strategies, and materialized views, cutting latency by 20%. Skills: PL/SQL · oracl · Java