Harsh Shah

Software Engineer at Microsoft

Integrated centralized fraud checks into container deployment scripts to prevent misuse; built dashboards to monitor zombie containers and optimize deployments; upgraded security to certificate-based bearer tokens; decommissioned legacy ACI SingleTenantSM clusters, sunsetting obsolete clusters and saving the organization $2.3M annually; migrated ACI from Pilot Fisht to Asgard Warp; led the design and rollout of the Agent Upgrade Pipeline; developed Agentic AI to analyze Kusto logs for root-cause insights in incidents.

Software Engineer at Sumo Logic

Updated PostgreSQL receiver to scrape more metrics for the OpenTelemetry collector; built and maintained features for the collector for logs, events and metrics on Kubernetes clusters; contributed to open-source efforts to optimize go-ps packaging.

Software Engineer at ActionIQ

Built integrations with Google, Salesforce, Tray.io, TikTok, and Facebook Ads to deliver user data to platforms using Java, Spring Boot, and REST APIs; implemented scheduled exports with Luigi; designed ProfileAPI to capture real-time data in MongoDB via Kafka Streams and route to AWS Lambda; developed Journey 2.0 with Redshift, Python, Luigi, and TypeScript; led Delta export writer feature reducing export time from 7 hours to 20 minutes; improved stage deployment costs and CI/CD; deployed Activation internal API on AWS ECS Fargate; built Datadog metrics; won 3 hackathons; implemented tests and fixed bugs during on-call.

Software Engineer at Modular Mining Systems

Implemented auto-interaction and efficient path algorithm for Autonomous Haulage Truck (AHT) at the mine; built REST APIs in Spring to enable third-party communication with Frontrunner server after token acquisition via WebAppHost in C#; developed TFSAnalyzer analytics on the TFS board to prioritize bug fixes, automated tests, and feature development; created a React/Kotlin JS frontend to automate MaxFleet maintenance workflows; contributed to the automation framework for the AHS team and created automated tests with Java and Cucumber.

Project at Question Answer System (Smart assistant) (Self-initiated)

Developed a smart assistant that answers questions across categories (movies, songs, geography) by querying a SQL database; used Python with Spacy/NLTK/WordNet; applied dependency parsing and ML to classify question types and generate SQL queries.

Project at Hadoop Map-Reduce on AWS (Self-initiated)

Implemented Map-Reduce using Hadoop to map connections among 4.5 million papers; deployed on AWS EMR; XML parsing in Java and mappers in Scala; visualized results with Gephi.

Project at Smart City Bike Share (Self-initiated)

IoT-based bike-sharing solution; deployed Python app on Google Cloud to monitor bike location and health from on-bike sensors; created REST API in Node.js to detect threats based on sensor readings and updated Firebase.

Project at Visualization of Opioid Abuse (Self-initiated)

Led a data visualization project with domain experts to tell stories about opioid abuse using D3.js; ranked 1st in class among 30 teams.

Project at Git Application Analysis (Self-initiated)

Simulated an enterprise-level DevOps environment in Linux containers using Java; automated Jenkins DSL pipeline to create and update projects, including static analysis, JaCoCo coverage, and Redis-backed issue storage.

Software Engineer at Caterpillar Lab, UIC Innovation Center

Configured PL-542 in Linux to store accelerometer data in CSV and analyzed trends to predict machine activity; used LSTM to forecast machine actions at construction sites; added GPS tracking to PL-542 to improve ML model accuracy.

Project at Auto Grader (Self-initiated)

Developed an automatic scoring system for ETS essays focusing on length, spelling, grammar, and coherence using NLP tools.