I build fullstack applications, and I care about owning them end-to-end. From planning how something ships, to making sure it's built the right way (clean architecture, proper testing, CI/CD, whatever the project actually needs).
I'm pretty great at debugging and root cause analysis. I'm especially good at taking control over someone else's code and figuring out how it actually works.

Resume

My Work & Experience

My journey through technology and innovation

Sept 2025 - Dec 2025

Visionify Inc.

Software Developer Intern

    Worked on Pallet Vision, an analytics application tracking production across multiple pallet manufacturing facilities. Diagnosed and fixed a critical performance bottleneck, cutting configuration save latency by 94% (~17s down to under 1s) by decoupling blocking backend processes. Migrated legacy authentication to SSO, refactoring the auth flow and middleware for better security and reliability. Built automated alerting pipelines for live production data, including inactivity detection and scheduled health reports. Integrated webhook-based data exchange with business partners for real-time updates. Wrote SQL-based preprocessing functions to turn raw edge-server telemetry into structured, analyzable data.
View Certificate
Next.jsTypeScriptTailwind CSSPostgreSQLGitAnt DesignSupabaseClerk (Auth)Vercel
May 2025 - Aug 2025

Shiransh Ventures Pvt. Ltd.

Web Developer Intern

    Built and deployed an end-to-end client-facing preschool web platform (React, TypeScript, Vite, Tailwind CSS), owning the project from requirements to production. Designed a multi-section admissions workflow with form validation and document handling, streamlining student onboarding end-to-end. Integrated Firebase (Auth, Firestore, Cloud Storage) for secure enquiry submission, data persistence, and file uploads across a responsive multi-page app with reusable components.
View Certificate
React.jsTypeScriptTailwind CSSFirebaseGit

My Projects

July 2026

Multi-Agent AI Research Assistant

Engineered a 4-agent research pipeline (research, reader, writer, critique) orchestrated with LangGraph, coordinating hand-offs through a shared state graph. Built a critique agent that cross-checks generated claims against retrieved sources and triggers targeted revisions through a bounded feedback loop, reducing unsupported claims in the final draft. Engineered a production-grade FastAPI backend with structured logging, custom exception handling, API-key authentication, and rate limiting. Defined a research endpoint to stream live per-agent progress via Server-Sent Events, consumed by a Next.js frontend. Containerized the backend with Docker and automated builds/deployments to Docker Hub and Railway via a GitHub Actions CI/CD pipeline.

PythonLangGraphTavilyBeautiful Soup 4FastAPIDockerGitGitHub ActionsRailwayNext.jsTypeScriptTailwind CSSVercel
April 2026

LLM Observability Platform

An observability platform that monitors LLM token consumption across applications using a custom npm wrapper library. Tracks usage metrics across multiple dimensions- per feature, per user, and per provider- enabling granular cost and performance analysis. Implements time-series token usage dashboards, giving teams visibility into usage trends and anomalies over time. Mirrors core infrastructure monitoring principles: instrumentation, metric collection, multi-dimensional aggregation, and reporting.

Next.jsTypeScriptTailwind CSSClerk (Auth)SupabaseSQLnpm

Education & Certifications

My academic journey and professional qualifications

Education

NMAM Institute of Technology

Bachelor of Technology

Information Science & Engineering

2026

GPA: 8.92/10

Certifications

Full Stack Web Development

100xdevs

Let's Connect!

Have an exciting opportunity for me?

Connect with me on LinkedIn or mail me at avyaysujir1010@gmail.com