Sandeep N

Senior Full Stack Software Engineer with 10+ years building and deploying production-grade REST APIs and scalable backend systems using Python (Django, FastAPI) and AWS, Azure Cloud, with hands-on React/TypeScript frontend experience. AWS Certified Solutions Architect – Associate, with deep expertise in serverless architecture (Lambda, API Gateway, CloudFormation, DynamoDB) and relational/NoSQL databases. Experienced building and deploying production AI agents and MCP servers integrating Claude and OpenAI models, including real-time, multi-channel AI-assisted systems with full-stack ownership.

Software Developer (Middleware, Python, Django, Java Restful) at Comcast (2022-05 – Present)

Client: Comcast, Philadelphia, PA. Project 1: Enterprise AI Support & Integration Assistant(Full Stack) Environment: Python, FastAPI, AsyncIO, Pydantic, Pydantic_AI, GPT-4o, Claude Sonnet models, GitHub Copilot, Claude Code, MCP, FastMCP, AWS Lambda, ECS Fargate, API Gateway, Application Load Balancer (ALB), ECR, S3, IAM, YAML, SSM Parameter Store, CloudWatch, Kinesis, GitHub Actions CICD, Shell/PowerShell DeepEval, Docker, CloudWatch Logs & Metrics, Web UI Chatbot, Slack Bot, React, TypeScript, Vite, CSS, react-markdown, TanStack Query (React Query), Vault. Project 2: Partner Integration API Platform

• Designed and developed an enterprise AI assistant end to end from backend agent architecture to the production chat UI.
• Developed an AWS Lambda-based AI agent leveraging OpenAI GPT-4o mini/ Claude Sonnet models via PydanticAI, integrating with custom toolsets and external MCP server endpoints for dynamic platform support automation.
• Architected the MCP-based tool orchestration approach for the AI assistant evaluating static vs. dynamic tool discovery and designed a FastMCP-based framework that let the LLM select the right tool at runtime across APIs, logs, and documentation systems, improving extensibility as new integrations were added without redeploying the agent's core logic.
• Built and deployed a FastAPI-based MCP server on AWS ECS Fargate for AI tool and prompt requests.
• Designed session-aware conversational workflows with contextual follow-up support, custom prompt engineering, and a slash-command framework to improve AI output relevance and efficiency.
• Extended the AI assistant across a Web UI chatbot and Slack bot, delivering multi-channel support to several partners.
• Automated full-stack infrastructure deployment (ECR, ECS Fargate, ALB, API Gateway, Lambda) using AWS CloudFormation and GitHub Actions CI/CD, with AWS SAM and SSM Parameter Store for multi-environment support.
• Built developer tooling and local environment setup scripts in Bash Shell, enabling repeatable local development stack initialization, AWS SSM parameter loading, and subnet querying designed for onboarding reliability and minimal manual intervention.
• Built the production chat UI in React and TypeScript (Vite, CSS, Framer Motion), including message list, input, conversation history sidebar, and slash-command prompt UX.
• Integrated react-query for server-state management and caching, with message-history sync and trace-ID propagation to preserve conversation continuity across sessions.
• Implemented secure markdown rendering (react-markdown + dompurify) to safely display AI-generated responses while preventing XSS, plus debounced inputs, loading/error states, and a global ErrorBoundary for resilience.
• Used Claude Code as a primary development agent for Python/FastAPI and React feature implementation, structuring tasks and SKILL.md-style specs to drive accurate, context-aligned code generation.
• Designed and implemented LLM evaluation tests using DeepEval, covering tool selection correctness, response quality, prompt behavior, and security/confidentiality across test modules.
• Led team-wide adoption of Claude Code alongside GitHub Copilot, establishing prompt and task-decomposition conventions that improved AI-generated code quality and reduced review cycles.
• Integrated Github Copilot into CI/CD workflows for automated code review and pre-merge validation, complementing SonarQube/Snyk gates to catch correctness and security issues in AI and human-generated code.
• Designed and developed RESTful APIs and microservice

Master’s in computer science/information technology – Campbellsville University

Bachelors in Electronics and Communication Engineering – JNTUK University