Hero background
Abhishek Yadav

Abhishek Yadav

CSE Student • Fullstack Developer • AI Systems Builder

GH in @
abhi740000@gmail.com

I'm Abhishek Yadav, a Computer Science and Engineering student at UPES focused on fullstack development and applied AI systems.
I build production-ready web applications, backend APIs, admin dashboards, and real-time inference pipelines
with an emphasis on reliability, security, and practical deployment.

Fullstack Frontend, backend, infra-minded implementation
Applied AI Inference pipelines, NLP systems, computer vision work

Skills

Python Python
JavaScript JavaScript
React React
Next.js Next.js
Flutter Flutter
Tailwind CSS Tailwind CSS
Node.js Node.js
Express Express
FastAPI FastAPI
PostgreSQL PostgreSQL
MongoDB MongoDB
MySQL MySQL
Flask Flask
PyTorch PyTorch
LangChain
RAG
Agentic AI
MCP
Docker Docker
Git Git

Professional Experience

Naiyo24

FullStack Intern

June – Sept 2025

Flutter, React, PostgreSQL

  • Contributed to the development of cross-platform dashboards by integrating frontend components with backend APIs.
  • Worked on API integration and data handling for accounting workflows, ensuring correct data flow between frontend and backend.
  • Implemented secure session handling and validation for transaction-related features.

Freelance

Fullstack Developer

Node.js, Express, React, PostgreSQL

  • Designed and delivered end-to-end web applications including backend APIs, authentication systems, and database architecture.
  • Built admin dashboards and internal tools enabling non-technical users to manage data without direct database interaction.
  • Implemented secure authentication, role-based access control, and third-party integrations.

Projects

ZaFlora

Built the commerce stack end-to-end for a live perfume brand, covering storefront experience, authentication, checkout, and admin workflows with a focus on secure payments and dependable operations.

End-to-End Built catalog, checkout, auth, and admin flows in one system.
Live Payments Integrated Razorpay with reliable transaction handling.
  • Designed scalable backend APIs and database flows for products, users, and orders.
  • Implemented secure authentication and role-aware admin workflows.
  • Shipped a production-ready storefront connected to real operational tooling.
Node.js Express Supabase better-auth Razorpay
Live Site
TruthLens

Real-Time Fake News Detection Pipeline

Developed an AI-assisted misinformation analysis system that combines model inference, entity-aware NLP, and multi-source validation to support fast, higher-confidence decisions.

95% Model accuracy achieved with a custom-trained LSTM pipeline.
Real-Time Inference deployed for live prediction rather than offline-only analysis.
  • Combined spaCy NER, semantic similarity, and source validation in one workflow.
  • Containerized inference with Docker and deployed the model for responsive predictions.
  • Focused on practical signal quality instead of one-shot classification alone.
Flask Next.js PyTorch NLP Docker HuggingFace

Building a Low-Latency Video Intelligence Pipeline

Engineered a stream-to-inference system for live computer vision workloads, optimized around sustained throughput, low latency, and stable detection output across continuous video feeds.

Low Latency Kept frame streaming and inference non-blocking across the pipeline.
Live Detection Rendered real-time predictions with optimized ONNX-based object detection.
  • Used WebSockets for continuous frame transfer without breaking interaction flow.
  • Integrated YOLO, ONNX Runtime, and OpenCV for efficient detection and rendering.
  • Architected the pipeline to reduce frame loss under sustained live input.
FastAPI WebSockets ONNX Runtime OpenCV YOLO

Education

B.Tech in Computer Science and Engineering (Fullstack AI)

2022 – 2026

University of Petroleum and Energy Studies (UPES), Dehradun

Achievements