Hi, I'm Priyabrata ๐Ÿ‘‹
AI/Agent Engineer at QAD Inc. I build LLM-powered agentic workflows for enterprise supply chain automation. Check out my blog
Bengaluru, Karnataka, India
PM

About

I'm an AI/Agent Engineer with 1 year of experience building production-grade LLM applications and Agentic Systems for ERP and supply chain automation. Currently, I work at QAD, where I focus on designing scalable AI agents, evaluation frameworks, observability systems, and workflow automation solutions for enterprise customers.

I hold a B.Tech in Computer Science from Techno Main Salt Lake (CGPA 8.55), where I ranked 2nd in the department. Outside of work, I enjoy competitive programming and have solved 2500+ problems across platforms, achieving Guardian on LeetCode (2178) and 4โ˜… on CodeChef (1917).

Work Experience

Q

QAD Inc.

Dec 2025 - Present
AI/Agent Engineer
  • Improved the reliability and production readiness of ERP automation agents (PO creation, production order closure, and file mapping) by resolving critical workflow, validation, and data integrity issues for enterprise QAD Adaptive ERP customers.
  • Optimized the file mapping agent by introducing on-demand schema retrieval, reducing token consumption by over 80% per batch and improving scalability for complex data mappings.
  • Developed automated agent evaluation and regression testing frameworks to validate workflow accuracy, tool usage, and response quality before production deployments.
  • Replaced Lambda-based invocations with direct HTTP APIs for alert creation, improving cross-environment reliability (dev, preprod, prod) and eliminating deployment blockers.
Technologies:PythonLangChainLiteLLMJQFastAPIAWS LambdaPydanticasyncio
K

Kavida AI (Acquired by QAD)

Aug 2025 - Nov 2025
Junior Data Scientist
  • Led the redesign of Agent Maxi's core agent architecture to support large-scale enterprise customers, replacing LangChain chains with a native Python + LiteLLM implementation. Enabled real-time streaming SSE for multi-agent reasoning, improved observability, and significantly reduced framework overhead to support production-scale workloads.
  • Implemented Langfuse observability for end-to-end LLM tracing across multi-agent workflows and integrated SigNoz for route-level API monitoring and production observability.
  • Redesigned the architecture of three business-critical monitoring services, consolidating duplicated logic, optimizing database connectivity through memoized resource initialization, and reducing failure points to improve performance and operational stability.
  • Developed AI-powered news article scraping pipeline for supplier risk intelligence.
Technologies:PythonLiteLLMLangfuseSigNozFastAPISSEasynciouv
G

Grouple

Apr 2025 - Jun 2025
Full Stack Developer Intern
  • Revamped the homepage using Next.js and Tailwind CSS, significantly improving load time and UI responsiveness.
  • Fixed critical bugs in the email service responsible for end-user communication and improved navigation and user interaction flow.
  • Deployed the website to a custom domain via AWS EC2 with a CI/CD pipeline, and configured NGINX for efficient serving with caching, compression, and routing rules.
  • Worked with AWS RDS for database management, optimized queries, and cleaned the database by removing redundant entries and dropping unused tables.
Technologies:Next.jsTailwind CSSNode.jsAWS EC2NGINXAWS RDSCI/CD

My Skills

Python
TypeScript
JavaScript
Go
C++

My Projects

LaResume

ATS-friendly resume builder with LaTeX export and PDF generation. 2900+ signups, 4000+ resumes created, 190+ GitHub stars.

Next.jsTypeScriptAWS EC2DockerMongoDBPrisma

Quotica

Full-stack AI image generation app using OpenAI and Gemini APIs with context-aware chat and webhook-based payments.

Next.jsTypeScriptOpenAIGeminiMongoDBPrisma

StackOverflow Tags Predictor

ML model that predicts tags for Stack Overflow questions, simplifying categorization and discovery.

PythonJupyterFlaskJavaScriptDockerAWS

Tasks CLI

Lightweight CLI task manager built with Golang to track to-dos directly from the terminal.

Go

Currency Converter TUI

Terminal-based currency converter with a rich TUI built in Golang, with real-time exchange rates. Installable via Homebrew.

Go

Get in Touch

Please contact me directly at priyabrata8558@gmail.com or through this form.

Want to read my thoughts on tech? Check out my blog on Medium โ†’