Overview:
Ashish Agarwall is the accomplished Head of DevOps and Architecture at BCS Technologies, bringing nearly a decade of transformative experience in the DevOps and cloud infrastructure landscape. With a unique journey from electronics engineering to .NET development and ultimately to DevOps leadership, he has guided over 100 professionals across industries through complex digital transformations. His expertise spans enterprise-scale DevOps implementations, hybrid cloud strategies, and the integration of cutting-edge AI into software delivery pipelines.
What We Talk About:
- Ashish’s remarkable transition from electronics engineering through .NET development to becoming a DevOps leader during the pivotal 2017 cloud revolution, showcasing the evolution of modern software delivery practices.
- The reality of enterprise infrastructure decisions reveals why most organizations maintain a 70% on-premise, 30% cloud split driven by security, compliance, and economic considerations rather than pure technical preferences.
- A practical philosophy for DevOps success: “If you have good DevOps, you have good sleep. If you don’t, you’re working 20+ hours a day” – emphasizing human-centered outcomes over tool-centric approaches.
- Real-world transformation stories, including dramatic improvements from 8-hour manual deployments with frequent failures to automated, zero-downtime releases through strategic implementation of DORA metrics and DevOps best practices.
- The pragmatic integration of AI in DevOps workflows, from intelligent code analysis and automated PR reviews to proactive log analysis and autonomous issue resolution, is balanced with the critical need to maintain core engineering skills.
Podcast Highlights
DevOps Transformation Philosophy:
- DevOps success is measured by quality of life: good implementation enables restful sleep, poor implementation creates 20+ hour workdays
- Breaking down traditional silos between development and operations teams through shared ownership models
- Emphasis on cultural transformation alongside technical implementation
- “Production is everyone’s responsibility from day one” – fostering collective accountability for system reliability
Hybrid Cloud Reality:
- Enterprise preference for 70% on-premise, 30% cloud infrastructure, driven by practical constraints
- Banking and financial services face regulatory barriers preventing full cloud migration of sensitive data
- Security, compliance, and data sovereignty requirements often outweigh pure technical considerations
- Cost analysis reveals cloud isn’t always more economical, especially for high-volume data operations
DORA Metrics Implementation:
- Deployment Frequency: Transformation example from 12 deployments per year to 48+ deployments annually
- Lead Time for Changes: Reducing time from code commit to production through automation and testing
- Change Failure Rate: Target of zero failures through comprehensive pre-production validation
- Mean Time to Recovery: Building self-healing systems with rapid incident response capabilities
AI Integration Strategy:
- Development-time assistance: Real-time security analysis, code quality checks, and automated refactoring in IDEs
- Automated code review: AI-powered PR analysis providing instant feedback before human review
- Intelligent monitoring: Log noise reduction and actionable insight extraction from system telemetry
- Proactive issue resolution: Agentic systems handling repetitive infrastructure problems autonomously
Configuration Management Lessons:
- A three-day production outage caused by scattered configurations across multiple systems
- Best practice: Limit configuration to one or two centralized, source-controlled locations
- Infrastructure as Code ensures consistency across development, testing, and production environments
- Predictable infrastructure patterns enable confident application promotion through the deployment pipeline
Common DevOps Anti-Patterns:
- Manual deployment processes lasting 8+ hours with frequent rollbacks and hotfixes
- Environment inconsistencies are causing production-specific failures
- Complex branching strategies create integration and deployment complexity
- Reactive monitoring leading to after-the-fact problem discovery
Future-Ready Engineering Advice:
- Technology cycles now change every 6 months to 1 year, versus the previous 10-year stability
- Recommendation: Become expert-level in 1-2 areas while maintaining broad technology familiarity
- Balanced AI usage: Leverage for productivity but maintain core skill development through hands-on practice
- Full-stack competency is increasingly essential for understanding end-to-end system behavior
Key Takeaway:
Successful DevOps transformation requires balancing technological capabilities with organizational culture change, regulatory constraints, and economic realities. The most effective approach focuses on human outcomes (team well-being, reduced stress) rather than purely technical metrics, while strategically adopting automation and AI to handle repetitive tasks and enable teams to focus on high-value creative problem-solving. True DevOps maturity is achieved when reliable, automated systems allow engineering teams to sleep well at night, confident that their infrastructure can handle both planned deployments and unexpected incidents with minimal human intervention.