Skip to content
View JayDS22's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report JayDS22

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
JayDS22/README.md

Jay Guwalani

Typing SVG

I build enterprise-scale distributed systems, real-time inference pipelines, and production software. My work spans 24B+ records/day at sub-200ms latency, with $122.9M+ in quantifiable cost savings through ML-driven optimization.

LinkedIn Portfolio Email ResearchGate


Featured Projects

Domain Project Highlights
GenAI / LLM Production LLM Serving Framework vLLM continuous batching, 12.3K req/sec, 42ms P50, INT4 quantization, 70% memory savings
Multi-Agent AI ConstellationAI Multi-Agentic RAG LangGraph orchestration, semantic caching, real-time token optimization, GPU monitoring
Data Engineering Enterprise Data Warehouse 24B records, <60s latency, 100K records/sec, 32 fact tables, 128 dimensions
Distributed Systems Big Data Query Platform on K8s Distributed query engine on Kubernetes, horizontal scaling, fault-tolerant execution
Medical Imaging Real-Time Medical Image Enhancement DDPM + 3D U-Net for CT/MRI, SSIM: 0.89, <2s on 512³ volumes, 92% radiologist approval
Quantitative Finance Quant Trading Platform LSTM/Transformer HFT, Sharpe: 2.1, 87.3% directional accuracy, 1M+ ticks/sec
Computer Vision Industrial Quality Control CV YOLOv8 + ResNet-50, 94.2% accuracy, 500+ images/min, <150ms, TensorRT
Infrastructure Automated Infra Provisioning Terraform, Ansible, Puppet, VMware vSphere/NSX-T, 85% faster deployments, 95% compliance
Experimentation Real-Time Experimentation Platform Thompson sampling, CUPED, causal inference, Benjamini-Hochberg, 50+ experiments

See all 70+ repositories →


What I Work With

Languages · Python · Java · Scala · SQL · TypeScript · JavaScript · C++ · Go · R

ML & AI — PyTorch · TensorFlow · Scikit-Learn · HuggingFace · LangChain · and more

PyTorch · TensorFlow · Scikit-Learn · XGBoost · LightGBM · HuggingFace Transformers · LangChain · LangGraph · OpenAI · Anthropic Claude · MLflow · Weights & Biases · Optuna · Ray Tune · ONNX · TensorRT · vLLM · NVIDIA NeMo · Stan · PyMC

Backend & APIs — FastAPI · Flask · Spring Boot · Node.js · and more

FastAPI · Flask · Django · Spring Boot · Node.js · Express · REST · gRPC · GraphQL · WebSockets · Celery · RabbitMQ · API Gateway

Data & Distributed Systems — Spark · Kafka · Airflow · Snowflake · and more

Apache Spark · Apache Kafka · Apache Airflow · Prefect · Apache Flink · Delta Lake · Apache Hudi · Databricks · Snowflake · dbt · Apache Hive · Presto/Trino

Databases & Storage — PostgreSQL · MongoDB · Redis · Neo4j · and more

PostgreSQL · MySQL · MongoDB · Redis · Elasticsearch · Apache Cassandra · Neo4j · ClickHouse · DynamoDB · Redshift · BigQuery · Pinecone · ChromaDB · S3 · HDFS

Cloud & Infrastructure — AWS · Azure · GCP · Kubernetes · Terraform · and more

AWS: SageMaker · Bedrock · Lambda · Step Functions · EKS · ECR · EventBridge · Glue · Kinesis · Redshift · S3 · CloudFormation · CDK Azure: Data Factory · Synapse · Purview · AKS · Azure ML GCP: Vertex AI · BigQuery · GKE · Dataflow · Pub/Sub IaC & Containers: Terraform · Ansible · Puppet · Docker · Kubernetes · Helm

DevOps & Monitoring — GitHub Actions · Jenkins · Prometheus · Grafana · and more

GitHub Actions · Jenkins · ArgoCD · GitOps · Prometheus · Grafana · ELK Stack · Datadog · CloudWatch · PagerDuty · SonarQube · Trivy · CI/CD pipeline design

Analytics & Visualization — Tableau · Power BI · Plotly · Streamlit · and more

Tableau · Power BI · Apache Superset · Looker · Plotly · Streamlit · Jupyter · Matplotlib · Seaborn · D3.js

Statistical Methods — Bayesian Inference · Causal Inference · Survival Analysis · and more

Bayesian Inference (Stan/PyMC) · MCMC · Hierarchical Models · Causal Inference · Propensity Score Matching · Difference-in-Differences · Instrumental Variables · Survival Analysis (Cox/Kaplan-Meier) · Time Series (ARIMA/GARCH/State-Space) · A/B Testing · Multi-Armed Bandits · Sequential Testing · Power Analysis


Impact

24B+
records/day processed
<200ms
real-time inference
$122.9M+
ML-driven savings
13.4K
crashes prevented
99%+
system uptime
74.4%
infra cost reduction
3
research publications
130+
professionals mentored

Recognition

Best Performer Hackathon Judge Research

  • 🏆 Best Performer FY 2023-24 — Bridgestone Group, recognized for enterprise AI delivery
  • ⚖️ Hackathon Judge & Mentor — HackMIT, UCLA LA Hacks
  • 🎤 Invited Speaker — ML conferences & technical advisory boards
  • 📚 3 Research Publications — IEEE, IJAET & viXra in predictive maintenance & ML

Publications

  • Predictive Maintenance in Automotive Telematics using Machine LearningviXra:2511.0028 ML models (K-NN, Random Forest, SVM) for truck load classification via CAN bus telematics; 95.97% accuracy
  • IEEE & IJAET — Predictive maintenance algorithms with 64.53% accuracy improvement across 300K+ vehicles

GitHub Streak

Pinned Loading

  1. ConstellationAI-Multi-Agentic-RAG ConstellationAI-Multi-Agentic-RAG Public

    Production-ready multi-agent RAG system with LangGraph orchestration, real-time token optimization, GPU monitoring, semantic caching, and comprehensive performance analytics. Built for scale.

    Python 2

  2. Autogen-Retrieve-Chat-System Autogen-Retrieve-Chat-System Public

    Production-ready RAG system using AutoGen's multi-agent framework for intelligent document analysis, automated code generation, and complex question answering with dynamic context management.

    Python

  3. Real-Time-Experimentation-Platform Real-Time-Experimentation-Platform Public

    🧪 A/B testing platform with Thompson sampling multi-armed bandits achieving 23% higher conversion rates 📊 Sequential testing framework with Benjamini-Hochberg correction, causal inference, and Baye…

    Python

  4. Insurance_Modelling Insurance_Modelling Public

    This repository contains a ML Project utilizing Linear Regression Models, GLMs, and various methodlogies to perform indepth analysis of a Motor Insurnace Data.

    Jupyter Notebook

  5. Multi-Robot-Coordination-Framework Multi-Robot-Coordination-Framework Public

    Multi-Robot Coordination Framework with distributed reinforcement learning achieving 92% convergence, <50ms allocation, and 99.9% availability Scalable autonomous robot fleet coordination using Q-l…

    Python 1

  6. Multi-GPU-Distributed-Training-Framework Multi-GPU-Distributed-Training-Framework Public

    Production-ready multi-GPU distributed training framework with DDP/FSDP, gradient compression, and 89% scaling efficiency at 16 GPUs. Includes TensorBoard monitoring, auto-checkpointing, and Kubern…

    Python 1 1