jayant pahuja

Purpose: Lead the development of AI-driven learning content solutions, collaborating with leadership on product strategy and proactively researching and implementing cutting-edge machine learning techniques.

• Automated content analysis and mapping for learning materials (books, presentations, articles), reducing manual effort by 80%. This system enabled efficient topic selection and client discussions.
• Developed an Agentic system to automate book adaptation from existing learning materials with human in the loop.
• Created a testing environment ("playground") to optimize AI agent performance for continuous improvement of our content adaptation system.

Purpose: Designed, developed, and implemented innovative solutions for a global ad campaign management system, leveraging machine learning and data strategies to optimize efficiency and enhance user experience.

• Developed a data-driven billboard selection algorithm to speed up campaign creation & evaluation process.
• Built a robust testing framework enabling objective evaluation and faster iteration of the billboard selection algorithm. Framework was later adopted for use in other company products.
• Streamlined client contracting & campaign management across multiple teams through automation, resulting in improved communication, reduced errors, and more accurate financial projections
• Developed natural language search for ad campaign management platform, enabling faster frame discovery through complex queries like "frames in London near subways in November".

Purpose: Leveraged machine learning and data engineering expertise to develop and optimize a BI product for out-of-home advertising, delivering high-impact user segmentation and exposure measurement solutions.

• Lead adoption of geospatial models to process anonymized location data (GDPR compliant via PII deletion), improving report generation time by 70% without significant loss of accuracy.
• Developed geospatial model predicting bus ad audience exposure, enabling out-of-home advertisers to optimize campaigns and target demographics.
• Engineered KPIs for early detection of data drift (e.g., location shifts, distribution changes) in geospatial models, reducing debugging time and minimizing downtime.
• Implemented a data visualization and exploration platform, enabling rapid feedback on model outputs and early user involvement in data validation, reducing model iteration cycle time.