Aman Chandra

GENOME AI Decision Assistant: Architected and developed a policy-grounded Agentic AI assistant using LangGraph and RAG, enabling customer-specific decisions by combining account data with authoritative business policies. Implemented a hybrid retrieval pipeline combining BM25 and Qdrant to retrieve relevant policy evidence using keyword-based and semantic search, improving contextual relevance and grounding LLM responses. Designed an Agentic decision workflow with specialized nodes and tools for RAWI: Built the core of RAWI, a React 18 + TypeScript app for exploring Qatar’s 2025 Census.

It combines a Mapbox map of municipalities and zones, an AI chat where users ask questions in plain language, and dynamic analytics charts (bar, donut, heatmap, year-over-year comparison) built with Recharts and Redux. Built the AI chat so one message can hold several questions, with question history, session reset and PDF export of chosen answers. The chat stays in sync with the map, so it zooms and fil

FullStack AI Engineer at PiCode Systems (2026-05 – Present)

• GENOME AI Decision Assistant: Architected and developed a policy-grounded Agentic AI assistant using LangGraph and RAG, enabling customer-specific decisions by combining account data with authoritative business policies. Implemented a hybrid retrieval pipeline combining BM25 and Qdrant to retrieve relevant policy evidence using keyword-based and semantic search, improving contextual relevance and grounding LLM responses. Designed an Agentic decision workflow with specialized nodes and tools for customer eligibility assessment, policy compliance, decision explanations, and automated service-request routing. Developed evidence-backed responses with policy citations, document versions, section references, and account-level evidence to improve decision transparency and explainability. Developed a FastAPI backend and React interface to demonstrate end-to-end interactions, decision outcomes, supporting evidence, and service-request confirmations. Validated agent behavior against 15+ business scenarios covering policy eligibility, vulnerable customers, disputes, and service-request routing.
• RAWI: Built the core of RAWI, a React 18 + TypeScript app for exploring Qatar’s 2025 Census. It combines a Mapbox map of municipalities and zones, an AI chat where users ask questions in plain language, and dynamic analytics charts (bar, donut, heatmap, year-over-year comparison) built with Recharts and Redux. Built the AI chat so one message can hold several questions, with question history, session reset and PDF export of chosen answers. The chat stays in sync with the map, so it zooms and filters to the municipality or zone each question is about. I also packaged the chatbot as a separate embeddable app (a guarded iframe route) so other census apps can use it. Made the app fully bilingual (English and Arabic), including right-to-left charts, maps, numbers and PDF export, used i18next to made it more reliable.

Senior Software Engineer at LTIMindtree (2022-07 – 2026-04)

• AI-powered Architect Workbench: Led Project development using React.js and Redux Toolkit, enabling artifact lifecycle management, approvals, and version control. Built and integrated AI Copilot chatbot and multimodal inference supporting PDFs, PPTs, images, and videos. Implemented a Retrieval-Augmented Generation (RAG) pipeline using LangChain, HuggingFace embeddings, and semantic similarity search to retrieve relevant document chunks and generate context-grounded LLM responses. Architected reusable API and state management layers using Redux Toolkit, React hooks, and memoization to minimize unnecessary re-renders and improve scalability and delivered MVP within strict timelines.
• AI Agent Marketplace: Developed a full-stack AI Agents Marketplace using React.js, Typescript, Node.js, Express.js, and MongoDB, enabling users to discover, browse, and interact with available AI agents. Built responsive marketplace interfaces with reusable components, supporting agent listings, search, filtering, and detailed agent views. Designed and integrated RESTful APIs using Node.js and Express.js for agent retrieval, agent management, and search/filter operations.

Integrated

MongoDB with Mongoose for schema-based agent data modeling and efficient querying of agent information.

• Temenos Assist – Generative AI Conversational Platform: Developed LLM-powered conversational assistant with dual-model playground for model comparison and runtime parameter tuning. Implemented secure token lifecycle management with auto-refresh. Built feedback system with MongoDB integration and integrated Speech Recognition API enabling voice-based AI interaction. Reduced bundle size by 35% and duplicate development effort by 40%
• Optimized frontend performance using code splitting, lazy loading, memoization, and efficient rendering techniques, improving application load time by 20%+ and enhancing Core Web Vitals.

B.Tech in Electrical Engineering – Indian Institute of Technology Ropar (2018-07 – 2022-05)