Available for opportunities

Shouvik
Choudhury

Agentic AI Engineer & AWS Specialist

AI Systems Engineer specializing in agentic workflows and autonomous orchestration. ~4 years building intelligent data pipelines and scalable LLM-driven applications — designing self-healing systems that leverage RAG, tool-use, and multi-agent frameworks.

Get in touch View work

Years Exp.

9.70

CGPA

30%

Throughput ↑

Professional Experience

Where I've
made an impact.

AWS Engineer

Cognizant

2022 — Present

Architected scalable data ingestion frameworks using Python & AWS to feed high-velocity data into LLM-based agentic workflows, improving processing throughput by 30%.
Engineered robust AI-ready data pipelines on AWS using Airflow, optimizing infrastructure for low-latency retrieval required by autonomous agents and intelligent orchestration layers.
Built automated data refinement pipelines that reduced latency by 40%, ensuring real-time data availability for AI models and agentic decision-making systems.
Developed and maintained automated workflows across AWS S3, EC2, and Glue — reducing manual intervention and accelerating delivery timelines by 25%.
Orchestrated multi-agent systems to bridge siloed data ecosystems, enabling autonomous cross-platform communication and shared intelligence.
Developed self-healing data pipelines that leverage LLM-driven agents to detect, categorize, and auto-correct quality issues without human intervention.
Designed and consumed RESTful Python APIs to integrate data sources, automate service communication, and expose pipeline outputs as endpoints — enabling seamless interoperability across microservices and external platforms.
Implemented monitoring and alerting systems for data pipelines using AWS CloudWatch and custom Python scripts, proactively identifying and resolving issues to maintain 99.9% uptime.
Collaborated with cross-functional teams to design and implement data models that support LLM-based applications, ensuring efficient storage and retrieval of structured and unstructured data for agentic AI workflows.
Conducted code reviews and provided mentorship to junior engineers, fostering a culture of best practices in data engineering and cloud architecture within the team.

Core Competencies

Tech I work
with every day.

🤖

Agentic AI & Gen AI

LLM Orchestration LangChain LangGraph RAG Vector DBs Autonomous Agents Prompt Engineering

☁️

Cloud & Big Data

AWS S3 / EC2 / Glue PySpark Apache Airflow FastAPI Terraform

📊

Analytics & Data

ETL Development Pipeline Automation Data Modelling Data Quality Monitoring & Alerting SQL / DynamoDB

🛠️

Libraries & Tools

Python Pandas NumPy TensorFlow Git / GitHub

Projects

Things I've
built.

ShantiView: AI Mental Health Companion

Python · Gen AI · FastAPI · NLP · Computer Vision

AI-driven mental health platform leveraging multimodal inputs and LLMs to assess well-being and deliver personalized, empathetic support insights.

Live Demo

Sacred Texts RAG: Multi-Religion Knowledge Base

Python · Gen AI · FastAPI · RAG · NVIDIA NIM

RAG-based knowledge system that retrieves and synthesizes insights from sacred texts, delivering context-aware answers with precise scriptural citations.

Live Demo

RoamWise: AI Travel Companion

Python · FastAPI · Gen AI · NVIDIA LLM · REST APIs

AI-powered travel platform that generates personalized itineraries, budgets, and travel insights using LLMs, with support for chat-based assistance and image-based landmark recognition.

Live Demo

RepoGraph: Cross-Repository Impact Analyzer

AWS Neptune · OpenSearch Vector DB · LLMs · Mermaid.js · AWS Bedrock

An intelligent knowledge base that maps cross-repository dependencies by combining vector embeddings with graph databases. It enables developers to perform multi-repo impact analysis via a chat interface and visualize architectural relationships through automated diagram generation.

Sutra IQ: Multi-Agent AI Research Assistant

LangGraph · FastAPI · React · LLMs · Tavily Search · Vector Embeddings

An intelligent research system where specialized AI agents collaborate to produce comprehensive reports. The multi-agent architecture orchestrates planning, web search, analysis, and writing tasks through a unified interface, enabling in-depth research synthesis with streaming real-time updates.