Hi, I'm

Sri Shriya Challapuram

Forward Deployed ✦ AI Engineer

Embedding within enterprise customer environments to architect, evaluate, and deploy production-grade multi-agent systems. I design and build AI systems that move from research & prototypes to real-world tools with measurable impact, strong evaluation, and practical safety.

Portrait of Sri Shriya Challapuram

About me

Designing AI systems that are useful, reliable, and grounded in real-world deployments

AI research and Engineer working at the intersection of machine learning, forward deployed, system design, and real-world deployment. Over the last 3.5+ years, I’ve built AI-powered products, participated in hackathons, and worked on applied research projects that connect model experimentation with practical engineering. My work focuses on LLM systems, retrieval pipelines, evaluation, and building trustworthy AI tools that teams can actually use. I enjoy turning ambiguous ideas into deployable systems and bridging the gap between research thinking and product execution.

MS in Computer Science & Information Systems

George Mason University · Aug 2024 – Dec 2025

Coursework: AI, NLP, Component Based Software Development, AI and LLMs, Network Security

BS in Computer Science

Gitam University · Jun 2017 – Jun 2021

Coursework: AI, Data Structures, Database Systems, OS, Networks, ML

Skills

Agentic AI & LLM Engineering

  • LangGraph
  • LangChain
  • Frontier Labs AI Models
  • Prompt Engineering
  • Multi-Agent Cycles
  • Memory Mechanisms
  • Context Window Optimization
  • Hugging Face Transformers
  • Supervised Fine-Tuning (SFT)
  • Reinforcement learning from human feedback (RLHF)
  • Reinforcement Learning from AI Feedback (RLAIF)

Evaluation & Trustworthy AI

  • LLM Evaluation Harnesses
  • LLM-as-a-Judge
  • Eval Dataset Design
  • Failure Mode Analysis
  • Regression Testing
  • LangSmith
  • LangFuse
  • Regression Gates
  • Boundary Validation
  • Heuristic Validators
  • Guardrail Pipelines

Context Retrieval & Ingestion

  • RAG Architecture
  • Vector Embeddings
  • Vector Search
  • AST & Code Parsing
  • Few-Shot Prompting
  • Zero-Shot Prompting
  • Agent Patterns

Backend & Distributed Systems

  • Python
  • FastAPI
  • Java
  • TypeScript
  • JavaScript
  • SQL
  • Shell Scripting
  • React
  • Django
  • Node.js
  • Express.js
  • REST APIs
  • GraphQL
  • Pydantic
  • Supabase
  • PostgreSQL
  • MySQL
  • MongoDB
  • DynamoDB
  • Firebase

Cloud, Infrastructure & QA

  • AWS (EC2, S3, SQS, Lambda, CloudWatch, IAM, API Gateway)
  • Docker
  • Kubernetes
  • Serverless Architecture
  • Jenkins
  • CI/CD Pipelines
  • Boundary Value Analysis
  • Equivalence Partitioning
  • Zero-Rollback QA
  • Selenium
  • Playwright
  • JMeter
  • Postman
  • Jest
  • TestRail
  • BrowserStack

Client-Facing & Forward Deployed

  • Technical Discovery
  • Requirements Translation
  • Architecture Consultation
  • Technical Advisory
  • Scoping High-Stakes Problems
  • Workshops & Code Reviews
  • Cross-Functional Coordination
  • Stakeholder Communication
  • Bug Lifecycle Management
  • Git
  • Linux
  • GitHub Actions
  • Jira

Projects

Live App

Reduced developer onboarding from 3 months to < 60 minutes

AI-Powered Code Architecture & Visualization

Architected a full-stack codebase comprehension tool that parses unstructured source code into a dynamic dependency graph and generates semantic documentation using Claude via LangGraph, with context optimization and hallucination reduction.

PythonLangGraphFastAPIReactTypeScriptGraphQLAnthropic ClaudeNetworkXGraph VisualizationContext Optimization
Live App

80% reduction in manual planning time for early users

Autonomous Multi-API Travel & Scheduling Agent

Designed an autonomous event-driven workflow engine integrating multiple disparate enterprise APIs (Eventbrite, Google Calendar) to resolve dynamic scheduling conflicts without standardized schemas.

Claude APIn8nVanilla JSLeafletSupabaseREST APIs
Hackathon / Live Demo

Automated candidate resurfacing & unstructured scorecard synthesis

Autonomous Salesforce Agentforce Platform for Talent Rediscovery

Built at the Salesforce x MetroStar Hackathon. Designed two autonomous Agentforce agents ('The Recaller' & 'The Briefer') to automatically resurface qualified silver-medalist candidates and generate multi-round interview briefing dossiers.

Salesforce AgentforcePrompt BuilderApexSalesforce FlowsLLM Reasoning

Kubernetes Deployment Orchestrator

Web-based orchestration tool with React frontend and Python backend. Uses AWS Lambda, S3, and SQS. Streamlines multi-container Kubernetes deployments. Reduced deployment errors by 25% and improved resource utilization by 15%. Supports auto-rollback.

ReactPythonAWS LambdaS3SQSKubernetes

Image Captioning using Deep Learning

CNN–LSTM based image captioning system. Feature extraction via CNNs and sequence modeling via LSTM. Django-based web interface for uploads and caption display.

CNNLSTMDjangoDeep Learning

Multiplayer Drawing Game

Real-time collaborative drawing and chat. WebSockets, Node.js, JavaScript. Low-latency synchronization across users.

WebSocketsNode.jsJavaScriptReal-time

Your days, in bloom

Designed a gentle habit-tracking concept where daily goals become small promises to yourself. Phool turns progress into a growing bloom, making consistency feel visual, calm, and rewarding.

Figma MakeProduct DesignHabit Tracking

Experience

Get in Touch

I'm actively open to new opportunities, interesting projects, or just having a conversation about technology. Open to Forward Deployed AI Engineer & Production AI roles (open to travel and relocation)

Sri Shriya Challapuram // 2026