Apr 2024 – Apr 2025 · 1 yr

• Sports Pricing Quant at IMGArena/SportCast/Openbet
• Developed new dynamic AFL Player Goal Groups markets/prices, Football Player Score or Assist markets
• Implemented bivariate Poisson like simulation for draw prices to match industry draw prices
• Optimized trader variables for totals, quarter,3-quarter, half, whole handicap market lines
• Goalkeeper saves pricing based on Shots on Target and Blocked Percentage
• Resolved simulation issues with Goal kicks, Shots off Target, etc.,
• Pricing to prevent ARB opportunities on NFL acca
• Odds Conversion (Football, Basketball, American Football etc) (Python)
• Implemented 2-way, 3-way analytical, n-way optimization solver methods (differential margin - odds ratio, profit ratio etc.,) to remove margin from various odds and convert it to true probabilities across markets such as 1x2, total goals O/U, corners O/U, Asian handicap x.5, Double chance etc.,
• ML model for predicting corner/card expectations from over/under markets

Aug 2022 – Apr 2024 · 1 yr 8 mos

• Developed a high-quality Python application and optimized performance for the Open 2023 Championship (Golf) with concurrency enhancements on Apache Druid and managed historical data ingestion
• Conducted extensive EDA and ETL for various data sources, including football and golf historical data
• Achieved competitive Golf Player Ratings and Rankings Model and values players across all Golf Tournaments. Results are unique and yet comparable with Official World Golf Rankings (upto 0.8 Recall and upto 0.6 KTau)
• Trained Golf Hole Difficulty model with a Score - 0.75 along with designing mathematical formulation and demonstrated Unsupervised Live player influence indicator for MLS
• Led the optimization of Spark streaming jobs for reduced latencies and achieved significant cost savings on AWS by using spot instances and redesigning S3 API calls for data ingestion. Troubleshot class path issues for Spark-operator on AWS EKS, improving resource efficiency by 50%
• Achieved cost savings of $120,000 USD and optimized massive matrices for Golf Player Ratings
• Implemented efficient Live Major League Soccer data ingestion along with optimizing Druid queries for improved performance
• Fine-tuned Apache Spark Streaming for reduced latency

Jan 2022 – Jul 2022 · 6 mos · Leeds, England, United Kingdom · On-site

• https://bet.me/ - Sports Betting Exchange
• Created a hyperparameter tuning framework using Bayesian Optimization, reduced search time from hours to seconds
• Designed solutions for MLFlow on Kubernetes, incorporated in-memory high-speed messaging infrastructure and Python Async IO
• Led MLOps efforts for data and model pipeline development, customized MLFlow for tracking experiments, model training, tuning on Kubernetes, and model serving
• Evaluated forecast models using metrics like Brier Score, Ignorance Score, and Skill Score
• Conducted statistical tests to identify correlations between prices, betting operators' efficiency, seasonality, trends, and expected goals (xG) along with analysing the impact of managerial changes
• Developed and implemented novel score metrics for assessing football team skill levels and a master algorithm for merging teams based on various criteria
• Built data engineering code to load large sports datasets from diverse sources, handling hundreds of gigabytes of data using Golang, gRPC, and protobuf
• Employed techniques like Poisson Regression and Dixon Coles GLM for benchmarking

Apr 2019 – Jan 2022 · 2 yrs 9 mos · Hyderabad, Telangana, India

• Architected green field consumer Journey Analytics product/multi-tenant platform with Streaming and microservices architecture
• Actively saved $10 million in TCO and $700,000 in development costs and influenced AI/ML solutions with enterprise-wide impact
• Developed AI/ML solutions for various applications, including a granted ML patent for File Transformation
• Enhanced Master Data Management with intelligent algorithms and architected a modern streaming platform for multi-tenancy, pre-processing 3 million records daily, and implemented distributed tracing using Splunk and Spring
• Led enterprise-level programs for technology evaluation, modernization, and AI/ML integration, aligning with industry standards, resulting in cost savings and on-time releases

Feb 2018 – Mar 2019 · 1 yr 1 mo · Greater Hyderabad Area

• Led a team of 15 individuals in the creation and management of a sophisticated real-time data integration project involving Salesforce Platform Events (Kafka-based) and Mulesoft and handled over 3 million transactions daily
• Provided innovative and suitable solutions for challenges such as record rollback, maintained First-In-First-Out (FIFO) order of transactions, and resolving race conditions in bidirectional data flows
• Developed a strong proficiency in Distributed Version Control Systems (DVCS), particularly Git, and adopted best practices for branching, merging, and pull request strategies

Feb 2015 – Feb 2018 · 3 yrs · Hyderabad, Telangana, India

• 1.Developed integration of user identities and roles between Oracle SaaS and Oracle PaaS clouds for federation.
• "Oracle SaaS (Fusion Application) applications and Oracle PaaS(custom applications) can be integrated for a simple SSO experience.The SaaS Fusion Applications form the IdentityProvider for PaaS applications. Additionally the Fusion Applications's enterprise roles can be migrated for allowing PaaS applications to reuse and enhance the role based access control.(RBAC)"
• 2.Developed rest-compatible secure UI components(Oracle ADF and Oracle JET) consumed by Oracle Applications.

Feb 2013 – Feb 2015 · 2 yrs · Hyderabad, Telangana, India

• 1.Written password encryption algorithms (DES standard) in pure Java
• 2.Written parser for PKCS 12 formatted trustStore/keystores to extract private keys and X.509 certificates.(https://docs.progress.com/bundle/datadirect-oracle-jdbc-60/page/Oracle-Wallet-SSL-authentication.html)
• 3.Troubleshot very tricky protocol (DB2,Oracle,TDS etc., ) issues in multi-thread scenarios , SSL issues
• 4.Troubleshot some of the toughest Java memory surge and performance issues using hprofiling,jmap,Java runtime options ,garbage collection log analysis
• 5.Led teams for JDBC driver releases and NO SQL prototypes

Jan 2008 – Feb 2013 · 5 yrs 1 mo · Hyderabad, Telangana, India

Aug 2005 – Jan 2008 · 2 yrs 5 mos · Hyderabad, Telangana, India