Hi, I'mHarshit Shiroiya

Building and shipping clinical software products using Angular, .NET Core, Azure, and GenAI tooling.

• Developed 3 clinical software products as micro-frontend Angular applications using Angular Signals for reactive state management, eliminating redundant change-detection cycles and improving rendering performance by 40%
• Built and deployed custom GenAI agents and internal Copilot-style assistants on Azure OpenAI + LangChain to automate clinical-study metadata generation, code scaffolding, and PR-review summarization; drove team-wide GitHub Copilot adoption with shared prompting standards and code-review guardrails
• Prototyped a retrieval-augmented generation (RAG) assistant over internal clinical documentation using Azure AI Search and embedding models, cutting average time-to-answer for study-configuration questions from hours to under 2 minutes
• Deployed containerized microservices on Azure Kubernetes Service (AKS) with Azure Container Registry, improving scalability 40% and reducing infrastructure costs by 35%; authored unit and integration tests and participated in weekly code reviews across 3 product squads
• Rolled out a distributed caching layer using Azure Cache for Redis with .NET Core (C#) and Entity Framework, cutting API latency by 50% under peak workloads for 2,000+ concurrent users
• Designed event-driven microservice communication via Azure Service Bus; instrumented observability with Azure Monitor and Application Insights and supported on-call rotations, reducing incident mean-time-to-detect (MTTD) by 60%
• Built a version-control system for clinical studies on Azure Blob Storage ensuring full audit traceability for regulatory review; introduced GraphQL alongside REST, cutting API response time by 35% and eliminating over-fetching

Built distributed data pipelines and NLP-based research tools for large-scale social media sentiment analysis.

• Built a distributed data-ingestion pipeline integrating Twitter, YouTube, and Reddit APIs to process high-volume sentiment data for 164 entities; optimized SQL workflows to cut processing time 30% and surface medication-efficacy insights across records for 10,000+ patients, enabling NLP-based emotion classification for 1,000+ researchers
• Introduced caching and async processing alongside Azure Data Lake Storage and Azure Service Bus for large-scale data management and reliable service communication, reducing latency by 50%
• Designed optimized MongoDB schemas for time-series social data with compound indexes, enabling sub-second querying across multi-million-document collections

Built scalable full-stack web applications with focus on accessibility, cloud deployment, and marketing automation workflows.

• Built scalable .NET Core (C#) backend microservices and WCAG 2.1-compliant frontends using Angular and Vue.js, integrating AWS S3 and Amazon RDS (PostgreSQL) to support a 30% scalability improvement across high-traffic marketing-automation workflows
• Configured a Memcached server-side caching layer that reduced database round-trips by 40% and boosted throughput for peak-load campaigns; automated CI/CD pipelines using AWS CodePipeline, cutting release cycle time by 25%