Chandra – AI agent orchestration, LLM, Python

A UK-based IT consultancy specializing in Enterprise AI Architecture and Engineering, focusing on transitioning AI from experimentation to production. The firm delivers, within a 10-week framework, high-value AI use cases, ranging from Proof of Concept to functional, multi-agent systems

• Provided technical leadership and solution design consulting for enterprise clients in regulated industries, including healthcare, financial services, and payments.
• Delivered end-to-end AI and cloud-native architecture engagements.
• Led integration design across complex multi-vendor ecosystems.
• Guided cross-functional delivery teams of 15+.
• Worked hands-on with agentic AI, RAG pipelines, LLM orchestration, and integration architecture across AWS, Azure, and GCP.
• Delivered deep technical design and stakeholder-facing advisory in high-compliance environments.

A leading international technology-led services and business process outsourcing company. Headquartered in the United Kingdom, it provides complex administration, regulated payments, and technology solutions to a vast global corporate client base.

• Led a multi-region AI modernization program, rearchitecting legacy IVR systems across the UK, US, and India using Amazon Connect, Lex, and Lambda to enable multilingual self-service and real-time call deflection.
• Designed and implemented an agentic AI framework using AWS Bedrock and RAG.
• Created autonomous agents capable of orchestrating complex journeys such as identity verification, fraud detection, and secure payments.

A public-facing GenAI chatbot embedded on the company's digital website aimed at answering questions about case studies, approach, and culture.

• Led the solution architecture for the platform, a public-facing GenAI chatbot embedded on their digital website.
• Worked with marketing, sales, and engineering to take the chatbot from strategy to first go-live in around seven weeks.
• Used existing website content, including 54+ case studies and company documents, as the initial knowledge base.

A voice-first automation platform for claims.A voice-first automation platform designed for managing claims.

• Delivered a voice-first automation platform for claims using Amazon Connect, Lex (ASR/NLU), and Bedrock.
• Wrote the core Python Lambda functions orchestrating the conversational flow.
• Developed a RAG-based knowledge lookup system to provide real-time fraud flag analysis and policy information to the voice agent.
• Reduced manual review by 30%.

A real-time video analysis platform, designed to elevate clients' insights and decision-making.

• Architected and developed a real-time video analysis platform.
• Used Google Colab notebooks to train a YOLOv8 model, which achieved more than 90% accuracy in brand recognition and passenger counting.
• Engineered a scalable data pipeline on AWS EKS.
• Contributed Python code for GPU-optimized container services for high-performance inference.

An internal Sales GPT platform using an agentic, RAG-based architecture to generate client proposals, delivery plans, and team structures from unstructured inputs and internal knowledge.

• Led the design and build of Ask ANDy, an internal Sales GPT platform.
• Engineered a modular RAG framework using n8n, Qdrant, Postgres, Redis, and Gemini.
• Built the framework so it could be reused across clients and domains by swapping data sources, access layers, and workflows.
• Turned Ask ANDy into a repeatable accelerator for Discovery and POCs.