Client: HK Infosoft

The Challenge

HK Infosoft’s corporate gifting operations depended on employees manually navigating a growing catalogue of products sourced from multiple vendors. Customer enquiries were typically expressed in natural language, such as “premium onboarding gifts under ₹2,500” or “executive gifts for banking clients”, but traditional keyword searches and spreadsheet filtering struggled to understand customer intent.

Finding the most relevant products often required employees to manually browse multiple categories, compare similar items, and verify pricing before recommendations could be shared with customers. This increased response times and made product recommendations inconsistent across the sales team.

– Traditional keyword searches frequently failed to identify relevant products because customer language differed from catalogue terminology.
– Employees manually reviewed hundreds of products before identifying suitable recommendations.
Budget compliance had to be verified manually before proposals could be shared with customers.
– Product recommendations depended heavily on employee experience rather than intelligent search.
– HK Infosoft wanted to introduce AI-powered search without managing foundation models or dedicated AI infrastructure.

The primary objective was to build a semantic product discovery platform capable of understanding customer intent, delivering intelligent recommendations, and reducing manual effort while maintaining a secure, serverless AWS architecture.

Discovery

DevOps TechLab conducted a collaborative discovery workshop with HK Infosoft’s business and technical stakeholders to understand the existing product discovery process and identify opportunities for AI-driven automation.

– Analysed the existing corporate gifting catalogue and product metadata to understand search limitations.
– Identified semantic search as the highest-value capability for improving employee productivity and recommendation quality.
– Evaluated self-managed AI infrastructure against Amazon Bedrock and selected Amazon Bedrock to minimise operational complexity.
– Defined core platform capabilities including natural-language search, semantic ranking, budget-aware recommendations, and intelligent product filtering.
– Established security requirements including private networking, IAM-based access, runtime secret management, and elimination of third-party AI API keys.

The discovery phase produced a clear AWS-native solution architecture focused on semantic search, operational simplicity, and long-term scalability.

Onboarding

Based on the discovery findings, DevOps TechLab designed a fully serverless architecture that enabled HK Infosoft to modernise product discovery while maintaining a low operational footprint.

– Designed the solution using Amazon API Gateway, AWS Lambda, Amazon Aurora PostgreSQL Serverless v2, and Amazon Bedrock.
– Selected Amazon Titan Text Embeddings V2 to generate semantic vector embeddings for both catalogue products and customer search queries.
– Designed the vector search layer using pgvector with HNSW indexing inside Amazon Aurora PostgreSQL to deliver fast similarity searches.
– Reviewed the proposed architecture, security controls, and projected operating costs with HK Infosoft stakeholders before implementation.
– Planned a phased implementation approach to minimise business disruption while enabling rapid adoption.

This onboarding process ensured that business objectives, architecture decisions, and AWS security best practices were fully aligned before development began.

Operations & Support

DevOps TechLab implemented, deployed, and validated the AI-powered semantic product discovery platform within HK Infosoft’s AWS environment.

– Built an AI-powered natural-language search experience capable of understanding customer intent instead of relying on keyword matching.
– Generated vector embeddings for the complete product catalogue using Amazon Bedrock and indexed them with pgvector for high-performance semantic search.
– Implemented intelligent recommendation logic combining semantic relevance with business rules such as budget limits, vendor preference, and product availability.
– Deployed the React application using Amazon S3 and Amazon CloudFront with Amazon API Gateway and AWS Lambda providing a fully serverless backend.
– Secured the environment using AWS Secrets Manager, Amazon RDS Proxy, IAM roles, private subnets, and VPC endpoints for secure service communication.
– Provisioned the complete infrastructure using AWS CloudFormation, ensuring repeatable deployments and simplified environment management.

Each deployment stage was jointly validated with HK Infosoft to ensure functional accuracy, search quality, security, and operational readiness before production rollout.

Optimisation & Advisory

Following production deployment, DevOps TechLab worked with HK Infosoft to optimise search performance, infrastructure efficiency, and long-term scalability.

– Tuned Amazon Aurora Serverless v2 to automatically scale according to search demand, reducing idle infrastructure costs.
– Optimised vector similarity search using HNSW indexing to improve semantic search performance across the product catalogue.
– Implemented Amazon RDS Proxy to improve database connection management for concurrent AWS Lambda workloads.
– Recommended VPC interface endpoints in place of a NAT Gateway to reduce recurring networking costs while improving security.
– Defined a future roadmap including Amazon Cognito integration, automated credential rotation, enhanced analytics, and expanded AI-powered recommendation capabilities.

This optimisation and advisory engagement ensures HK Infosoft has a scalable, secure, and cost-efficient AI platform that can continue evolving alongside future business requirements.

Summary

By combining Amazon Bedrock, AWS Lambda, Amazon Aurora PostgreSQL Serverless v2, and a fully serverless AWS architecture, DevOps TechLab helped HK Infosoft transform product discovery from a manual catalogue search process into an intelligent semantic search experience.

Employees can now search using natural language, receive AI-powered product recommendations, and identify relevant products in seconds instead of manually reviewing large catalogues. The solution improves search accuracy, reduces operational effort, and provides a scalable AWS-native foundation for future AI innovation without requiring HK Infosoft to manage AI infrastructure.

About DevOps TechLab

DevOps TechLab helps organisations design, build, and operate cloud-native and AI-powered solutions on AWS. Our expertise spans serverless architectures, Amazon Bedrock, generative AI, data engineering, DevSecOps, and AWS Well-Architected best practices, enabling customers to modernise applications, accelerate innovation, and deliver secure, scalable cloud solutions with confidence.