Yash Chheda

Technical delivery owner for a 10-engineer platform team accountable for 100+ Kubernetes clusters across 25+ AWS accounts and 25 AWS regions, running both internal and customer workloads.

• Drove AIOps-based automation of operational toil, cutting manual operational interventions ~35% and reducing MTTR ~40% — capacity returned directly to roadmap delivery.
• Lead cloud cost optimization across the fleet, delivering ~$1M in annual savings by tying capacity planning to growth forecasts rather than reactive scaling.
• Upgraded 100+ clusters across 25 regions through 4 Kubernetes minor versions in 18 months with zero customer-facing downtime, using staged rollout and automated pre-flight validation.
• Moved fleet configuration onto GitOps reconciliation, replacing imperative cluster changes with declarative desired state that is reviewed before it lands and auditable after it does.
• Extended the software supply chain with SBOM generation, container image signing and policy-as-code admission control, so provenance is enforced at deploy time rather than attested after the fact.
• Sustain audit readiness across FedRAMP Moderate/High, SOC 2, HIPAA, PCI DSS and DoD IL5 through platform and OS hardening (CIS Level 1 benchmarks), FIPS 140-3 validated cryptography, and control evidence — 3 assessments a year with zero platform findings.
• Brought AI workloads onto the platform: GPU node pool topology, scheduling and autoscaling for inference serving, lifting GPU utilization from ~30% to ~65% by sizing against real model traffic instead of peak provisioning.
• Extended the audited boundary to cover those AI workloads — approved model inventory, retention rules for prompts and responses, and egress control on model calls — so AI features ship under the same FedRAMP and IL5 controls as everything else.
• Matured FinOps practice from account-level reporting to workload-level unit economics, holding commitment coverage at ~85% and an effective savings rate of ~30%, with spend shown back to owning teams.
• Raised platform delivery cadence from fortnightly to daily releases while holding change failure rate under 5%, owning delivery across quarterly cycles with product, SRE and security stakeholders.
• Grew the team to 10 engineers with 3 internal promotions and zero regrettable departures, and cut new-engineer time to first production change from 6 weeks to 3, while staying active in architecture and design review.

Technical lead, Agile squad coach and mentor for the cloud platform team through the multi-year Stratus migration.

• Served as one of the technical leads on the Stratus cloud migration, moving 400+ customer sites off legacy infrastructure onto Kubernetes-native infrastructure over ~3 years — a programme spanning multiple squads.
• Mentored 4 engineers through weekly 1:1s focused on professional growth, and contributed to staffing, capacity planning and onboarding across the platform organisation.
• Served as Agile squad coach — sprint planning, estimation, reviews and retrospectives — while remaining a hands-on contributor in Go and Terraform.
• Acted as escalation point for high-visibility production incidents, converting findings into automation and runbooks instead of tribal knowledge.
• Raised the technical bar on Go, Terraform and Kubernetes practice through design reviews, pairing and documentation that shortened new-engineer ramp time.

• Built the Kubernetes-native foundation for the Stratus platform, growing adoption to cover the majority of all Appian customer deployments.
• Developed Kubernetes operator modules in Go and provisioned AWS infrastructure as code with Terraform.
• Built AWS integrations to provision secure, scalable customer Kubernetes sites, and extended the cloud operations dashboard into the single entry point for every activity on a customer cloud site.
• Root-caused production failures where customer deployments would not start, turning each into a permanent platform fix rather than a one-off recovery.
• Isolated a Linux kernel defect that froze customer site volumes and drove it through AWS support to a fix shipped on the AWS side.
• Designed, shipped and documented Appian Web APIs, and led two engineers as secondary squad coach across sprint planning, estimation, reviews and retrospectives.

• Architected and delivered customer applications end to end on the Appian low-code platform in SAIL, from requirements through design, development, testing and maintenance.
• Cut a core business process from 28–30 minutes to 4–5 seconds by refactoring legacy logic and process models.
• Designed and built the mission-critical Appian Certification programme application.
• Replaced third-party plugins with equivalent in-product features written in Java, and developed Java-based availability checks for critical third-party integrations using RPA bots.
• Built nightly synchronisation between Appian and external systems including Salesforce and HRIS, and mentored new developers; received department Badge Awards for top performance and ticket resolution speed.

Full-stack engineer on a digital mortgage lending platform, retained through the company's acquisition by One American Bank.

• Built the microservices architecture for a fintech lending platform: .NET Core 2.1 and C# 7 Web APIs, Vue.js and Razor front end, MS SQL Server, deployed as Azure App Service containers.
• Designed the Azure API Management gateway, routing external third-party traffic through an internal conversation logging proxy to backend services.
• Implemented OAuth 2.0 authorization via IdentityServer4, validating access tokens on every request entering the gateway, and delivered single sign-on across the application suite.
• Created CI/CD build and release pipelines from scratch, managed containers with Docker Compose, and migrated an in-house queue implementation to RabbitMQ.
• Added test coverage with Coverlet, OpenAPI documentation via Swagger UI, and exception logging through Application Insights.
• Built Encompass SDK plugins to update loans, delivered an in-house eSign and document merge feature, and redesigned the Broker Portal UI.