HARSH KAUSHAL

VIEW

// Synthetic Pod Crash - detected by OWL

Platform Engineer experienced in designing and maintaining reliable cloud systems that support business-critical applications. Known for streamlining deployments, improving performance, and ensuring consistent production stability.

Projects

• Autonomous K8s Diagnostic Agent is an AI-driven Kubernetes controller that automates root-cause analysis for container failures. It actively monitors the cluster for issues and leverages LLMs to instantly provide actionable fixes. Read Here

• Serverless Infrastructure Health Scanner is a event-driven AWS system that monitors infrastructure health signals and automatically triggers diagnostics and alerts using Lambda, EventBridge, and CloudWatch. Read Here

• Wix Installer is an end-to-end installation package that reduces manual setup effort by bundling runtime, database, and server components into a single executable, ensuring reliable and repeatable deployments. Read Here

Work & Experience

I don’t just write code; I design systems to solve specific problems. For me, a project is successful when it's reliable, observable, and actually moves the needle.

HCLTech

VIEW

//Highly Available EKS Architecture

Managing the architectural evolution of production-grade Kubernetes platforms. My focus is on eliminating deployment friction through parallelized IaC and slashing operational overhead through automated FinOps—moving the infrastructure from a cost center to a high-velocity asset.

Execution & Results

• FinOps: Engineered an automated environment scheduler using AWS Lambda & EventBridge, reducing monthly cloud expenditure by 50% ($10K to $5K) for non-production environments.

• Platform Engineering: Architected a scalable EKS platform supporting 13+ microservices with 100+ RPS for critical services, implementing Karpenter for node scaling and KEDA for event-driven pod autoscaling.

• Infrastructure as Code: Reengineered Terraform into modular, dependency-optimized stacks to enable parallel provisioning, reducing deployment time by 50% (90 to 45 minutes).

• Reliability: Migrated high-traffic services from Lambda to EKS for sustained stability; implemented Elasticache layers and Pod Disruption Budgets to ensure high availability during failures.

• CI/CD Innovation: Integrated Liquibase via Helm pre-install hooks for atomic DB schema versioning; built custom entrypoint logic to automate heap dump pushes to S3 upon OOM/Segmentation faults.

• Observability: Established data-driven capacity planning by building Grafana dashboards for P95 latency and RPS trends, enabling precise baselining for microservice replicas.

Unmanned Ground Vehicle - DTU

VIEW

VIEW

// Abstract and Depth Estimation

Autonomous vehicles usually need expensive lasers (LIDAR) to see. We didn't have the budget for that. So, We designed a system that uses standard webcams and smart math to simulate those lasers. It worked so well that we published a research paper on it.

Execution & Results

• Leadership: Managed 40+ engineers across software, electronics, and mechanical domains.
• Operations: Handled budget allocations and component sourcing to keep the prototype on track.
• The Tech: Fused Lane Reconstruction with Depth Estimation to generate pseudo-LIDAR scans at 12.5 FPS.
• Validation: Published at the IEEE ICARA 2022 conference as a guide for low-cost self-driving setups. Read Here

IBM | Qiskit Global Summer School

Completed an intensive, two-week program centered on advanced quantum simulation and quantum chemistry. The focus was on moving past perfect mathematical simulations to programming real-world, error-prone quantum hardware.

Execution & Results

• Hardware Cost Mapping: Composed highly optimized quantum circuits, calculating operational costs to minimize gate penalties on NISQ devices.
• Noise Modeling: Programmed real hardware simulations factoring in quantum noise, decoherence, and state-measurement errors.
• Physics Simulation: Modeled a physical Quantum Spin-1/2 condensed matter system, executing complex Hamiltonian time-evolutions using Qiskit.
• Completion: Successfully resolved all graded lab challenges on advanced circuit depth and noise mitigation. View Labs

DRDO | Solid State Physics Lab

Contributed to classified defense instrumentation projects focused on environmental threat detection. The work involved bypassing traditional data ingestion limits by engineering algorithms to isolate mechanical stress waves and process telemetry in extreme conditions.

Execution & Results

• Signal Isolation: Developed algorithms to filter and extract clean waveform signatures from highly corrupted, low-amplitude data streams.
• Telemetry Processing: Modeled time-series data to map elastic stress wave propagations, enabling the detection of sudden structural failures and fracture developments.
• Edge Optimization: Streamlined continuous data ingestion pipelines to handle high-frequency sampling without bottlenecking compute resources.
• Mission Impact: Provided critical research and algorithmic telemetry to support an active national project centered on real-time geohazard early-warning systems.

Certifications

HashiCorp Certified Terraform Associate (004) Mar 2026

AWS Certified AI Practitioner Mar 2026

AWS Certified Cloud Practitioner Dec 2024