Designing Resilient Pipelines
Engineering AI Insights
I am a Master of Science in Information Technology graduate from Arizona State University (GPA: 4.0/4.0), specializing in scalable ETL/ELT pipelines, real-time IoT telematics, machine learning models, and automated data observability frameworks using Python, SQL, and advanced business intelligence analytics.
Deshraj Jogiya
Data & AI/ML Engineer
11
Production Pipelines
2300
Git Commits
98
ETL Reliability %
4
ASU MS IT GPA
Professional Experience & Technical Impact
3+ years of designing production-grade data pipelines, optimizing LLM/AI models, and building high-reliability cloud data architectures.
Teleoperation Data Collection Associate
Objectways Technologies LLC | Tempe, AZ
- Collected and validated 10,000+ high-quality teleoperation sensor data samples for AI/ML model training, improving dataset accuracy and consistency by 20%.
- Engineered scalable data pipelines using Python, Scala, PySpark, and Kubernetes to process massive robotics datasets, cutting data processing time by 30%.
- Developed responsive web dashboards using React JS, JavaScript, and Java Spring Boot to monitor 10,000+ concurrent telemetry streams, improving system latency visualization by 30%.
- Executed robotics dataset QA and statistical anomaly detection, lowering downstream ML model training error rates by 15%.
- Leveraged AI-powered developer tools (Cursor, GitHub Copilot) to refactor backend microservices, write automated unit/integration tests, and maintain technical documentation.
Applied Machine Learning Engineer
Technoid LLC | Piscataway, NJ
- Optimized GPT-4o mini models for candidate resume matching using OpenAI APIs, SQL, and PostgreSQL, raising recommendation accuracy by 25%.
- Restructured Supabase data sync layers with PostgreSQL backend RLS fixes, cutting real-time vector synchronization latency by 65%.
- Architected backend microservices and RESTful APIs using Core Java, Spring Boot, and Python endpoints, reducing data sync latency by 65%.
- Built dynamic, accessible UI components in React JS connected to Spring Boot backend services, conducting peer code reviews to ensure Object-Oriented design compliance.
- Established automated phased regression and UAT testing frameworks using PyTest, cutting model deployment errors across AWS and Azure cloud environments by 30%.
Data Analyst
Zifatech Solutions LLC | Milwaukee, WI
- Transformed SQL/Python-based ETL pipelines to automate sales insights reporting, cutting manual reporting effort by 70%.
- Migrated legacy database workflows to AWS Glue & S3, increasing data availability by 60% and streamlining integration with Snowflake and Power BI.
- Modernized Star Schema dimensional models in Snowflake using Great Expectations QA validations, securing a 98%+ data reliability standard.
- Built executive Power BI dashboards connected to Snowflake data warehouses, converting complex data streams into strategic business insights.
- Engineered serverless RESTful API data pipelines on AWS (Glue, S3, Lambda) with Apache Airflow orchestration to aggregate multi-source application data, accelerating ingestion speed by 60%.
Data Engineer & ML Research Assistant
Arizona State University | Tempe, AZ
- Architected a real-time streaming pipeline utilizing Node.js & MongoDB, sustaining 99.9% uptime for 5,000 concurrent users and optimizing data throughput by 35%.
- Constructed a behavioral content recommendation engine, refining algorithms weekly to drive a 12% increase in click-through rates and lower user churn.
- Developed an NLP chatbot analyzing user input trends, identifying top UI friction points and boosting positive user satisfaction by 15%.
- Integrated transformer NLP classifiers (BERT) into web applications, boosting user sentiment feature extraction accuracy by 15%.
- Centralized application telemetry into SQL databases, presenting architecture designs, schema maps, and visual decks directly to leadership.
AI-ML Analyst Apprentice
Jetson Infinity | Austin, TX
- Implemented Python-based data processing pipelines leveraging Pandas & NumPy (+3% robotic arm precision, +10% workflow efficiency).
- Automated robotic arm motion ETL pipeline with Python & SQL, improving data processing speed by 17% and enabling real-time anomaly detection.
- Developed modular Python, C#, and .NET data transformation scripts, improving application dataset cleaning efficiency by 17%.
- Leveraged AI-powered developer tools (GitHub Copilot, Cursor) to write, test, debug, and refactor production-grade microservices and transformation pipelines.
- Devised an internal knowledge base with Python tutorials and real-world case studies, accelerating new analyst onboarding by 25% across Unix/Linux environments.
Data Analyst
Kronic Keys | Junagadh, GJ
- Cleaned and prepared large financial datasets (500k+ rows) in PostgreSQL, reducing monthly reporting cycle times by 30%.
- Engineered SQL queries and Tableau executive dashboards, contributing to a 12% increase in customer acquisition.
- Conducted weekly data integrity audits across relational database tables, eliminating duplicate records and ensuring 100% data quality control.
- Designed relational database schemas and indexed foreign keys in PostgreSQL to accelerate analytics query performance.
- Documented data definitions, data dictionary assets, and reporting workflows for business stakeholders.
Relentless Execution: Portfolio Repositories
Filter through active production-ready implementations built with end-to-end automation, modeling, and custom dashboards.
Job Search CRM & AI Application Tailoring Center
Local job placement command center with FastAPI and SQLite. Features AI-powered profile matching, resume and cover letter tailoring, LinkedIn recruiter outreach generation, and Kanban application tracking.
CurioSync: Serverless Tech News & LinkedIn Publisher
Scheduled serverless news curation and publisher pipeline running on GitHub Actions cron workflows. Integrates Gemini LLM chains and LinkedIn APIs, secured by Fernet symmetric encryption.
Member Messages QA System: Semantic Search & NER-Based Question Answering
Interactive QA API answering natural-language questions about member data, ranking unstructured messages via SentenceTransformer embeddings and identifying subjects with spaCy NER.
TalentVenue EventIntel: Enterprise Data Platform & Predictive Intelligence
Enterprise cloud data platform connecting SQL Server, Azure ADLS Gen2, Snowflake, and Streamlit. Deployed a Random Forest model predicting contract cancellation risk with 86.20% accuracy, alongside automated financial reporting and reconciliation workflows.
Reactive Job Ingest Gateway
A non-blocking job-posting ingestion gateway built on Spring WebFlux and R2DBC, republishing ingested postings as a live Server-Sent-Events stream.
Job Search dbt Warehouse
A dimensional dbt project (staging views into dimension and fact tables) running against a live Databricks SQL warehouse, with real data-quality tests in CI.
Job Market Lakehouse
A PySpark/Scala Spark SQL pipeline over 60k synthetic postings, gated by real Great Expectations checks and orchestrated end-to-end by a real Airflow DAG.
Job Notes
A company-research notes app: Express + MongoDB backend, Angular frontend, both with real automated tests.
Interview Prep API
A real ASP.NET Core minimal API (C#) for interview-prep questions, backed by Entity Framework Core + SQLite.
Company Blocklist
A real PHP/CodeIgniter 4 app tracking companies to avoid in a job search, with real automated feature tests.
Interview Prep Mobile
A real React Native (Expo) app for browsing/practicing interview-prep questions against a companion ASP.NET Core API.
Company Search (HTMX)
A live-search page using HTMX for partial-page updates -- FastAPI + Jinja2, no client-side JS framework.
ArcGIS Commute Insights
A job-search tool geocoding addresses and pulling real ArcGIS basemap vector-tile data via the live ArcGIS APIs.
AI Model Observability & Fairness Audits
Audits machine learning models in production by running Kolmogorov-Smirnov (KS) tests to track input data drift and evaluating Disparate Impact for bias protection.
FinTech Credit Risk & Fraud Command Center
End-to-end transaction ingestion and credit evaluation pipeline. Integrates a Random Forest classifier predicting loan eligibility with real-time transactional fraud checks.
Sentence Transformers, Multi-Task Learning & LLM Fine-Tuning
Built a Sentence Transformer from a Small BERT backbone with TensorFlow Hub, extended it into a multi-task architecture with classification and sentiment heads, and later added real LoRA fine-tuning of a modern generative LLM on a free GPU.
Multi-State Land Use Emissions Analysis
Geospatial emissions intelligence pipeline processing daily land cover changes across 5 U.S. states. Computes CO₂ inventory and forecasts emissions with ~90% accuracy using Linear Regression and Random Forest.
Tax Anomaly Audit Compliance Engine
Relational general ledger audit pipeline modeling transaction values following Benford's Law distribution and executing Isolation Forest anomaly flags for audit risk.
Automated Daily Data Insights
Stateless serverless daily financial data ingestion pipeline calling yFinance, computing rolling Z-score anomalies, and committing reports automatically via GitHub Actions.
Sales Operations & Customer Segmentation
Star schema database pipeline modeling Pareto 80/20 sales distributions. Automates RFM scaling and K-Means clustering to classify customers into 4 target personas.
Clinical Trials Patient Outcomes Analysis
Clinical database pipeline tracking patient survival and blood pressure across 3 treatment arms. Computes Log-Rank test statistics and Kaplan-Meier curves.
Real-Time IoT Telematics & Predictive Maintenance
Processes high-frequency EV fleet battery and motor sensor telemetry in streaming micro-batches, detecting Z-score anomalies and estimating Remaining Useful Life (RUL).
AI-ML Data Science Simulation
Retail data pipeline automating daily sales ingestion from 5 U.S. branches. Centralizes operations in SQL and uses Linear Regression/Random Forest forecasting to reduce stockouts by 15%.
Extending STEM across ASL
Inclusive educational learning platform utilizing a Convolutional Neural Network (CNN) in TensorFlow/Keras to recognize American Sign Language gestures for 7 core STEM concepts.
Chef at Gathering
Collaborated on an event coordination and professional catering booking React application, contributing to a 50% reduction in coordination time.
Stay Aware of Branch
Collaborated on a parent-school engagement application using React Native and Android frameworks, contributing to progress-report and attendance-tracking features that boosted parent involvement by 40%.
Get your Token
Collaborated on a Web3 crypto marketplace frontend and administrative dashboard using Express.js and Angular for NFT and Penky token transactions, contributing to a system that handled 1,000+ transactions in the first month.
City Forums
Collaborated on an online community engagement and marketplace platform using CodeIgniter 3 and MySQL, contributing to $10,000 in classified ad revenue through targeted marketing strategies.
Make It Short
Collaborated on integrating a link shortening engine and analytics tracking into a CodeIgniter 4 app, contributing to 10,000+ shortened URLs generated through gamified, redeemable points.
Solid Object Detection & Identification using Image Processing
Developed a shape detection and identification pipeline using PyTorch and OpenCV, achieving 98.97% classification accuracy by combining a custom CNN with traditional contour geometry approximations.
Interactive Engineering Demos
Test and monitor pipeline diagnostics or execute analytical SQL queries directly in-browser to see how data decisions are made.
Core Technical Expertise
Proficient across modern data engineering tools, machine learning modeling, database modeling, and visual analytics platforms.
Generative AI, LLMs & Agentic Systems
- LangChain, LlamaIndex & Semantic Kernel
- MCP (Model Context Protocol) & Tool Calling
- ReAct Agent Workflows & Autonomous Agents
- Retrieval-Augmented Generation (RAG) & Fine-Tuning (LoRA/PEFT)
- Vector Databases (pgvector, ChromaDB, Supabase RLS)
- OpenAI APIs (GPT-4o mini) & Gemini LLM Integration
- LLMOps & Agent Evaluation (LangSmith, Opik, Langfuse)
Machine Learning & Deep Learning Frameworks
- PyTorch, Jax, TensorFlow & Keras Deep Learning
- Scikit-Learn Modeling (Random Forest, XGBoost, K-Means)
- Computer Vision & Pattern Recognition (OpenCV, YOLO CNNs)
- Statistical Drift Auditing (KS-Test, PSI) & MLflow MLOps
- Responsible AI Governance (FTC & EU AI Act Bias Auditing)
- Financial & Audit Anomaly Detection (Benford's Law, Isolation Forest)
- Hypothesis Testing & Z-Test A/B Experimentation
Data Engineering, Platforms & Cloud
- PySpark & Spark SQL Large-Scale Data Processing
- Databricks Platform & dbt Dimensional Transformations
- AWS Cloud Infrastructure (S3, Glue, Lambda, EC2)
- GCP (Google Cloud Platform, BigQuery, Vertex AI) & Azure (ADLS Gen2)
- Snowflake Cloud Data Warehousing & Star Schema Modeling
- Apache Airflow & Dagster Pipeline Orchestration
- Great Expectations Data Quality Rules & Governance
- Table Formats (Delta Lake, Apache Iceberg, Parquet)
Full Stack & Software Engineering
- Python Backend Microservices (FastAPI, Flask, Asyncio)
- Core Java, Spring Boot & Spring Reactive Engineering
- C# & .NET Framework Backend Services
- JavaScript, TypeScript & Node.js/Express APIs
- React JS, React Native Mobile & Angular Web Interfaces
- PHP & CodeIgniter 3/4 Application Platforms
- RESTful APIs, GraphQL & Microservices Architecture
- Data Structures, Algorithms & Object-Oriented System Design
DevOps, Tooling & Analytics
- Docker Containerization & Kubernetes Cluster Scaling
- Git & GitHub Actions CI/CD Automated Pipelines
- Terraform Infrastructure as Code (IaC) & Jenkins
- AI Developer Productivity Tools (Cursor, Claude Code, GitHub Copilot)
- Power BI (DAX Modeling, KPI Dashboards)
- Tableau (LOD Expressions, Executive Storytelling)
- ArcGIS Geospatial Emissions Analytics & Chart.js Visualizations
- Relational Databases (PostgreSQL, SQLite, SQL Server, MySQL)
Live Learning Logs (TIL)
Real-time learning feed pulled dynamically from my `dev-root-affinity` learning log repository via the GitHub API, showcasing daily coding updates, concepts, and algorithms.
Fetching latest learning entries from GitHub...
Let's Connect
Looking for a technical Data Engineer / ML Engineer who combines mathematical modeling with production-grade data pipelines? Get in touch via these channels or send a message directly.
