I'm Mayur —
an AI Engineer & Web Developer
who builds intelligent AI products from scratch to scale.

Built a lightweight ML system for a B2B sales team to prioritize inbound leads and surface actionable insights.

• Designed a scoring model that ranked leads using behavioural data, historical conversions, and website interaction patterns.
• Implemented sentiment + intent extraction on inbound messages to classify prospects (cold, warm, hot).
• Automated generation of short sales briefs summarizing each lead's needs, objections, and recommended outreach angle.
• Integrated directly into the client's CRM so the sales team received real-time prioritization cues.

Impact: Helped the team focus on high-intent leads first, improving outreach efficiency and reducing manual triaging time by 40%.

Built an ML system that transforms noisy OCR scans and receipts into clean, structured data.

• Engineered pipelines that extract structured fields from low-quality documents.
• Improved multiclass inquiry-classification accuracy to 85%+ using fine-tuned models.
• Automated review workflows, reducing repetitive manual effort by ~60%.

Impact: Faster operations, fewer errors, and a scalable automation backbone.

Designed an intelligent agent that interprets customer web submissions and generates accurate, context-aware recommendations.

• Extracted essential signals from long-form user input.
• Classified inquiries by urgency and sentiment to eliminate response delays.
• Suggested next-step actions so support teams could deliver more personalized engagement.

Impact: Quicker triage, more efficient support teams, and human-like customer experiences.

Developed a resilient ingestion suite capable of handling deeply nested datasets across enterprise environments.

• Architected Azure Data Factory pipelines for ingestion, transformation, and landing into ADLS Gen2.
• Strengthened pipelines so downstream analytics remained stable despite complex JSON structures.
• Removed throughput bottlenecks and optimized workloads to cut compute and storage costs.

Impact: A dependable, scalable foundation for analytics and reporting across the organization.

Built an automated framework that enforces and validates data quality across teams.

• Implemented a validation layer using Apache Spark and Great Expectations.
• Converted stakeholder insights into automated tests and standardized runbooks.
• Created clear documentation for rapid onboarding and fast extension of the framework.

Impact: Higher data reliability, reduced manual QA, and consistent standards across teams.