Hi, I'm Shaurya.

Bengaluru, India

• Delivered a recurring investment service using Airflow, GraphQL, and Temporal. Priority scheduling and targeted payment retries cut processing for 30,000 daily orders from 5–6 hours to 30 minutes, supported over $5M in monthly investments, and recovered 75% of payment failures.
• Designed and deployed a multi-agent support platform with Google ADK and FastMCP, connecting API and database tools across domains and reducing ticket escalations to engineers by 70%.
• Engineered a Go-based meeting intelligence service using S3 and SQS to process recordings from 1,000+ meetings per month and automate summaries and action items.
• Developed a Slack-integrated AI service with specialized subagents, role-based guardrails, and a GKE sandbox for algorithm testing. It lets product managers create GitHub PRs and helps engineers diagnose production failures through logs, metrics, and post-deployment checks.

Bengaluru, India

• Architected a client asset-transfer platform integrating holdings APIs, consent forms, and SFTP transaction feeds with Cloud Run fallback jobs and S3 storage, enabling over $146M in transfers within six months.
• Automated daily transaction reporting for 30 asset managers with Celery and Redis, generating files for approximately 50,000 transactions in 10 minutes and allowing failed jobs to be regenerated on demand.
• Implemented a Django first-investment flow with payment APIs, automated failure recovery, and distributed locking to prevent duplicate transactions, enabling $500K in investments within two months.

Bengaluru, India

• Developed a Python merchandise-allocation system using historical click-through rates and the Google Sheets API, replacing a formula-driven manual workflow and reducing end-to-end allocation time by 66%.