Samridhhi Sharma

AI Full Stack Engineer · Bangalore, India

Full-stack and GenAI engineer building production-grade systems — from RAG pipelines and AI agents to scalable APIs and real-time applications. Clean code, pragmatic trade-offs, ownership mindset.

Indore Institute of Science and Technology
Samridhhi Sharma

Experience

C
Software EngineerCapri AI · Remote
Jan 2025 – Present
AI AgentsRAG PipelineEvent-DrivenLangChain
  • Architected a production-grade AI Agent Builder enabling seamless integrations with Google Calendar, Slack, and GoHighLevel for multi-provider enterprise orchestration.
  • Built an end-to-end RAG pipeline (ingestion → chunking → embeddings → vector search) powering a Knowledge Source module that trains AI agents on proprietary documents.
  • Integrated webhooks from multiple external providers into the core backend, enabling real-time event-driven workflows including automated meeting schedulers and inbound lead responders.
  • Delivered a Prompt Library feature for dynamic creation, management, and application of instructional prompts across bot configurations and scenarios.
  • Developed a full-stack document training interface with async job queuing and real-time user feedback for training status.
  • Stack: Python, FastAPI, LangChain, Node.js, Express, React, MongoDB, Pinecone, AWS
F
Software Development InternFreudia · Remote
Nov 2024 – Jan 2025
Multi-tenantSSORESTful APIs
  • Built the Find Therapists module enabling employees to search profiles and book sessions via real-time availability slots, improving care accessibility for enterprise users.
  • Developed RESTful APIs for New Organization Onboarding, automating multi-tenant URL provisioning and SSO-based access control via the SuperFreudia admin app.
  • Streamlined the client onboarding flow, reducing manual setup steps and accelerating feature configuration for new enterprise clients.

Skills

AI / ML
RAGLLMsVector DatabasesAI AgentsMCP
Cloud & DevOps
AWSDockerDockerNginxNginxJenkinsJenkinsGrafanaGrafana
Databases
PostgreSQLPostgreSQLMySQLMySQLMongoDBMongoDBDynamoDBDynamoDBRedisRedisFirebaseFirebase
Frameworks & Libraries
Node.jsNode.jsExpressExpressReactReactNext.jsNext.jsFastAPIFastAPILangChainLangChainPrismaPrismaJest
Languages
PythonPythonJavaScriptJavaScriptTypeScriptTypeScript
Tools
GitGitPostmanPostmanLinuxLinux

Projects

MedRAG

Clinical Q&A RAG System

Python, FastAPI, Node.js, PostgreSQL, pgvector, React, AWS, Docker

  • Built a production-ready RAG system for clinical Q&A delivering citation-backed LLM responses over large-scale medical document corpora.
  • Designed a scalable retrieval pipeline with semantic search via pgvector on PostgreSQL, enabling sub-100ms low-latency retrieval over high-dimensional embeddings.
  • Architected a modular backend with FastAPI and a Node.js/Express API gateway; delivered real-time streaming LLM responses with source attribution for a high-stakes medical domain.

AI Stock Prediction

AI-powered stock trend forecasting

Python, FastAPI, Prophet, Pandas, React, TypeScript, Docker, Nginx

  • Built a full-stack AI-powered stock prediction application using Facebook Prophet, enabling users to analyze trends and forecast future stock prices with machine learning.
  • Designed a high-performance FastAPI backend with RESTful endpoints for model inference, integrated with a React (Vite) frontend for interactive data visualization.
  • Containerized the entire application using Docker Compose with a multi-service setup; served the frontend via Nginx and deployed to cloud for public access.