Hi, I'm Shachi Mistry

Systems ML & Distributed Systems

Engineering high-throughput, low-latency machine learning systems, distributed training frameworks, and reliable cloud-native infrastructure.

bash - shachi@sjsu-core
shachi@sjsu-core:~$ cat profile.json
{
  "focus": "Systems ML & Cloud Infra",
  "education": "M.S. Software Engineering @ SJSU",
  "core_stack": ["Python", "C++", "Go", "PyTorch", "Kubernetes"],
  "status": "Building robust ModelOps platforms"
}
shachi@sjsu-core:~$

About Me

Systems background optimized for large-scale AI deployment.

I am a Software Engineer pursuing my Master's in Software Engineering at San Jose State University. My technical core is focused at the intersection of Systems Programming, Distributed Platforms, and Machine Learning Operations (MLOps).

I specialize in maximizing compute efficiency, managing GPU resources for heterogeneous environments, and building robust cloud-native systems processing multi-terabyte data workflows.

My goal is to advance the state-of-the-art in Big Model-as-a-Service (BMaaS) platforms, reducing the time-to-market for NLP, GenAI, and large-scale deep learning models worldwide.

Technical Engine

Languages

Python C++ Go Java SQL Bash

AI & Deep Learning

PyTorch NVIDIA Triton LangChain LangGraph Vector Search HuggingFace

Cloud & Distributed Systems

Kubernetes Docker AWS (SQS/DynamoDB) Apache Spark Kafka Airflow Redis

Projects Directory

Explore my open-source software projects and active telemetry simulations.

ML Platform

triton-inference-monitor

High-performance GPU inference benchmarking tool for NVIDIA Triton Server with automated latency profiling and hardware telemetry.

Python Triton NVML Docker
Generative AI

LayoverOS

Autonomous, state-persistent recovery agent for stranded travelers. Implements Vector search and stateful orchestration graph workflows.

Python LangGraph MongoDB Vector LLMs
Distributed Data

data-platform

Distributed data platform processing 5TB+ daily with 100+ concurrent jobs. Employs PySpark, Airflow scheduler, and Kubernetes pods.

Python Apache Spark Airflow Kubernetes
Cloud Architecture

scalable-order-processing-system

Cloud-native, event-driven order processing engine. Implements distributed idempotency keys, optimistic locking, and message queue patterns.

FastAPI AWS SQS DynamoDB Python
ETL & Streaming

Stock-Market-ETL-Analytics-Platform

End-to-end real-time ingestion pipeline processing stock market data streams using messaging clusters and validation triggers.

Apache Kafka Apache Spark Airflow Cassandra
Data Science

Netflix Popularity Prediction

Predictive analytics engine trained on historical Netflix catalogues, optimizing content rating predictions using feature engineering.

Jupyter Scikit-Learn Pandas XGBoost

Establish Contact

Reach out for collaborations, system designs, or technical inquiries.