Drashti Italiya
I’m Drashti Italiya, a software engineer with 5 years of experience building distributed services and production-ready LLM systems.
About
I’m Drashti Italiya, a software engineer with 5 years of experience building distributed services and production-ready LLM systems. I enjoy designing end-to-end solutions—whether that’s event-driven microservices, async task pipelines, or multi-tenant APIs backed by PostgreSQL at scale—while focusing on reliability, performance, and clean architecture.
At Adcore ON, I’ve helped cut major manual and operational bottlenecks by moving long-running workflows out of REST request paths, improving failure rates dramatically, and streaming job status back to clients in real time. I also build AI-enabled tooling (using pydantic-ai and Model Context Protocol) to support engineers and deflect repetitive support work, and I’ve delivered search and API performance improvements through migrations, indexing, and careful concurrency-safe backend work.
- 4 years' experience
- Joined September 2026
Skills· 12
Languages
- JavaneseIntermediate
Industry experience
- Software & Internet
- Financial Services
- Media & Entertainment
- Computers & Electronics
Work experience· 4 roles
Built an event-driven Python microservice for Media Blast to reduce weekly manual budget review from 6 hours to under 15 minutes. Used Google Cloud Pub/Sub to fan out state changes and RabbitMQ for direct service-to-service queues, keeping long-running analysis off the REST API path and streaming job status back to clients over WebSockets. Improved reliability for long-running media generation jobs by moving a synchronous endpoint to an asynchronous Celery pipeline backed by Redis, enforcing single-in-flight semantics to prevent stale-data overrides and pagination issues. Owned backend for Total Eclipse from empty repos to production launch, including tenant-isolated authentication, Google SSO, invitations, and session management across Docker containers on Linux. Shipped a redesigned Coupons onboarding flow with zero valid QA defects under a fixed deadline. Reduced schema-change review turnaround by building AI coding skills for Claude Code/Codex to validate Postgres migrations for tr
Developed a Python regression pipeline that pulled XML datasets from S3, validated records against expected schemas, and surfaced discrepancies in reporting views, reducing a recurring validation cycle from 24 hours to 30 minutes. Built an automated AWS distribution service (S3, Lambda, SNS) to deliver scheduled validation results to senior management without manual intervention. Implemented BDD smoke suites using Selenium WebDriver and Cucumber to catch regressions before release in a regulated environment and managed defect tracking to closure in qTest.
Reduced a recurring 24-hour manual validation cycle to 30 minutes by building a Python regression pipeline that pulled XML datasets from S3, validated records against expected schemas, and surfaced discrepancies in a reporting view. Built an automated AWS distribution service (S3, Lambda, SNS) to deliver scheduled validation results to senior management without manual effort. Added BDD smoke suites using Selenium WebDriver and Cucumber to catch regressions before release, with defects tracked to closure in qTest in a regulated environment.
Education
Certifications and awards· 10
Hire Drashti Italiya today
Post a job and invite Drashti, or contact them directly. No hidden commission.




