Jay Hemnani

AI enthusiast building intelligent automated systems.

Data pipelines, ML workflows, and production-ready solutions.

View Projects
25
Projects
66
Technologies
8
Domains

Career

Experience

Data Analyst

Summer 2025
  • Automated reporting pipelines with SQL/Python, reducing manual work and improving data consistency.
  • Built dashboards for occupancy and revenue tracking across multiple properties.
  • Developed demand forecasting models to support revenue management decisions.

Creative Lead

2022–2024
  • Delivered brand identity systems and marketing assets for clients in tech and hospitality.
  • Ran A/B experiments on campaigns and tracked engagement metrics.

AI/ML Intern

Jan–May 2022
  • Built credit fraud detection system using ML with techniques for handling class imbalance (SMOTE, ensemble methods).
  • Created data analytics dashboards with statistical models for operational reporting.

iOS App Development Intern

Jun–Aug 2021
  • Developed CodeLock, an iOS privacy/security app with encryption and secure local storage.
  • Tracked user engagement metrics and iterated on UX based on usage patterns.

SWE Intern

May–Nov 2019
  • Built a cloud-native communication platform for first responders using microservices on Azure.
  • Set up CI/CD pipelines for automated testing and deployment.

Background

Education & Research

B.Tech in Computer Engineering

Gandhinagar, India

GPA: 8.7/10
Software DevelopmentInternet of ThingsOperating SystemsBig Data Analytics+2 more

Publications

IEEE International Conference on Machine Learning & Data Science • 2022

Six-model stack achieving ~82.68% accuracy on PIMA dataset.

IEEE International Conference • 2021

Comparative analysis of CPU scheduling algorithms with priority queue optimization.

Selected Works

Projects

Builder

FIFA Soccer DS

Production-ready computer vision pipeline processing soccer videos at 22 FPS with real-time tracking, automated retraining, and REST API deployment.

PythonYOLOv8ByteTrackPyTorch GeometricFastAPI+2
Researcher

Soccer Vision Research

Next-gen sports analytics pipeline combining 4 SOTA models (RF-DETR, SAM2, SigLIP, ByteTrack) with modular architecture enabling rapid experimentation and benchmarking.

PythonPyTorchRF-DETRSAM2SigLIP+1
Builder

Nobel Data Intelligence

Production ML framework applying graph neural networks to molecular property prediction, processing 10K+ compounds with extensible pipeline integrating PyTorch Geometric and bioinformatics toolkits.

PythonPyTorch GeometricProDy+2
Builder

Website Watcher

Production web archiver monitoring 100+ sites every 2 hours with automated change detection, full-text search across 50K+ snapshots, and zero-downtime deployment.

PythonFlaskSQLite+3
Builder

CAG Deep Research

Production multi-agent research automation system built in 10 days, orchestrating 5+ specialized AI agents with LangGraph to autonomously synthesize 20-page research reports from raw queries.

PythonLangGraphLangChain+3
Builder

ContextBox

Personal knowledge assistant processing 1K+ screenshots with OCR, embedding 100K+ text chunks for semantic search, retrieving context in <200ms with vector similarity.

PythonLLMsOCR+3

Skills

My Toolkit

Technologies I've been building with lately.

Distributed Systems

Apache KafkaMicroservicesEvent-DrivenDockerKubernetes

AI & LLM Engineering

LangChainLangGraphAI AgentsRAGSemantic Search

Computer Vision

PyTorchRF-DETRSAM2SigLIPByteTrack

Data Engineering

Apache AirflowTimescaleDBETL PipelinesStream Processing

MLOps & Experimentation

Weights & BiasesMLflowExperiment TrackingModel Benchmarking

Full-Stack & Visualization

ReactNode.jsFastAPIStreamlitPlotly

Databases & Storage

PostgreSQLMongoDBRedispgvectorFAISS

Let's Build
Something

Currently open to full-time roles and select freelance projects.

© 2026 Jay Hemnani