Kunal Kaira

• Developed statistical models and spend simulations to reduce CAC from $6 to $1
• Standardized end to end AB experimentation practices, right from test design to post test readouts
• Designed and implemented self-serve initiatives leveraging self serve dashboards and Agentic AI that freed 50% of existing team capacity and enabled real time insights
• Led re-engineering of existing data pipelines reducing running cost by up to 40%
• Built in-house fraud detection model that leveraged social signal data to reduce promotion abuse by 45%

• Built and scaled cross-functional data teams (Analytics, ML, Data Engineering, BI) across the US, UK, India, Germany, and Israel
• Outlined vision and roadmap, subsequently leading and technically program managing execution efforts
• Defined hiring and performance rubrics as well as defined career pathing across all data work streams
• Led the development of a multi modal LLM based question answering service with internal routing mechanism as well as evaluation metrics to measure the completeness, informativeness, and abstraction of the generated answers, this service could answer
• AB Experimentation: Created best practices for Geo based as well as web/app-based AB experimentation
• Real time answer quality detection: Deep learning model to evaluate answer quality at the time of answer submission and provided diagnostic feedback to experts for needed correction. Helped reduce cost by $25 million per year while also leading to better user experience and retention
• Fraud Detection: Engineered multiple ML based solutions to detect account sharing and content abuse
• Smart Routing: Cosine similarity-based model(NLP) to route given question to the best live expert, this model reduced the TAT by 40% and improve answer quality by 20%
• Dynamic Pricing: Predictive and heuristic-based model to determine payout basis quality, TAT, question complexity as well seasonality
• High-Value Action: Persona-based model that identified high-value action to drive higher engagement and retention across customers; led to an ~10% increase in customer LTV
• Acquisition and retention forecast: Designed and setup an ensemble forecast model used to provide foundational information for organization level financial model
• Led development of multiple models across Marketing analytics like Mixed/Last touch attribution, MMM models, Lookalike audience targeting in FB/Adwords, Churn Models
• CRM (A/B Test, Propensity/Churn model-based targeting)- Email and Push Notification
• SEO: Co-led work on RSS feeds as well as catalog pruning for optimized SEO rankings
• Datamart creation to enable user segmentation, conversion and engagement deep dives
• Developed governance practices around data quality and reporting SLA's based on medallion architecture
• Discovery and finalization of tech-stack from both cost and evolution standpoint

Auto Insurance:

• Led a big data initiative that leveraged structured and unstructured call center data to quantify user satisfaction and enabling omnichannel experience
• Designed an enterprise-level claims solution that leveraged predictive analytics and accelerated visualization to improve claim TAT by 15%

Health Care:

• Developed machine learning model to flag members with high likelihood of making an avoidable ER for pre-existing conditions like asthma and diabetes, leading to cost savings of ~$6M
• Created unsupervised member, profile level clusters to identify cost saving opportunities by identifying outliers and creating highly customized and actionable intervention plans

Consumer Packaged Goods:

• Piloted a price elasticity model framework covering levels ranging from brand to the SKU Level, leading to increase in margin by ~18%
• Leveraged structured equation-based models (Bayesian lab) to identify brand claims driving brand equity; this work helped improve brand equity by 6% in the subsequent readout
• Built market mix modeling solutions enabling Brand and Portfolio level media optimization; this helped in driving marketing effectiveness by ~20%

• Developed time series model to provide market as well as sales rep level target/s
• Created overall ETL pipeline that provided end of day summary to sales rep across 38 states in US

• Co-developed financial product for ledger balance reconciliation across 20+ data formats
• Also, owned future implementation of change request to the reconciliation product