Featured Projects

Problem

Clients lacked real-time intelligence on market trends and inventory optimization in complex supply chains.

Solution

Built AI Agentic solution using Pydantic AI and LangChain, integrating with Snowflake for data analytics and AWS for scalable infrastructure.

Impact

• Real-time market insights
• Automated forecasting models
• Enhanced decision-making capabilities

Problem

Sales teams struggled to extract meaningful patterns from large volumes of seller data across multiple markets.

Solution

Created an AI-powered analytical system using advanced NLP and data processing pipelines to surface key insights automatically.

Impact

• Automated seller analysis
• Data-driven recommendations
• Reduced analysis time by 70%

Problem

Manual quote requests created bottlenecks and delayed response times for logistics partners.

Solution

Developed full-stack solution with automated matching algorithms and real-time communication systems.

Impact

• 50% faster quote processing
• Improved partner engagement
• Scalable B2B infrastructure

Problem

Legacy system took 45-60 minutes to generate text and audio forecasts, impacting operational decision-making and backlog of Meteorologists team.

Solution

Designed intelligent agent using AWS Bedrock and LangChain, optimizing data processing pipelines and inference models.

Impact

• 45 minutes → 20 seconds processing time
• 99.5% accuracy maintained
• Real-time forecasting generation capability

Problem

Healthcare provider spent excessive time manually reviewing and digitizing insurance claims.

Solution

Built automated pipeline combining Azure Document Intelligence, GPT-4, and custom validation logic for end-to-end processing.

Impact

• 75% reduction in manual review time
• Improved accuracy
• Scalable document processing

Problem

Manual payroll audits are time-consuming and prone to missing subtle patterns of irregularities.

Solution

Implemented machine learning models for anomaly detection with automated alert systems and reporting dashboards.

Impact

• Automated anomaly detection
• Early fraud identification
• Reduced audit time by 60%

Problem

Legacy monolithic system had performance bottlenecks and couldn't scale with growing transaction volume.

Solution

Implemented microservices-based solution in Go with event-driven architecture, optimized database queries, and comprehensive monitoring.

Impact

• 5x performance improvement
• 99.9% uptime
• £10M+ daily transaction processing
• Reduced latency by 60%

Problem

Slow build times, flaky tests and expensive cloud costs hindered development velocity and increased operational expenses.

Solution

Redesigned CI/CD workflows with parallel execution, intelligent caching, and infrastructure-as-code optimizations.

Impact

• 80% reduction in pipeline runtime
• £££k annual cost savings
• Improved developer experience

Problem

Manual order processing took 2+ hours, creating customer satisfaction issues and operational bottlenecks.

Solution

Developed cloud-native serverless architecture with automated workflows and real-time order tracking.

Impact

• 2 hours → 20 minutes fulfillment time
• 94% reduction in processing time
• Scalable architecture