The person behind the work

Engineering platforms.
Building teams.

I lead engineering for Concierto.cloud, the enterprise SaaS platform of Trianz (a $500M+ digital transformation firm). The organization is 200+ engineers across five product clusters - Migrate, Manage, Modernize, Insights, and Agentic AI - run through a layer of directors and engineering managers, one accountable owner per cluster, serving 80+ Fortune 500 clients at 99.9% platform uptime.

Running the organization

I scaled the group from 40 to 200+ engineers, and restructured it along the way from function-based teams into product-aligned Agile value streams, each cluster owning its own roadmap, quality bar, and production on-call. I run quarterly performance, calibration, and succession reviews across the management layer, and I have grown senior engineers into engineering-manager and architect roles so the structure stays self-sustaining as it grows. Hiring, onboarding, and retention practices we put in place cut attrition by 30% and halved time-to-productivity for new engineers.

I own the engineering headcount budget and the platform's cloud infrastructure spend across AWS, Azure, and GCP - accountable for both capacity planning and unit-cost reduction. I set technology strategy and the quarterly roadmap with Product, Sales, and Customer Success, tied to contracted client commitments. Shift-left quality practices took post-release defects down 60% and MTTR down 45%.

Platform and IP

I led the AWS acquisition of our data connector IP (Snowflake, BigQuery, IBM DB2, Oracle), now integrated into AWS Athena, working directly with AWS engineering through the code transition - and I'm currently engaged with the AWS CloudWatch team on security connectors. I adopted and extended Trino as our core analytics layer with custom connectors, query optimizations, and federation. Our backend pipelines process 10M+ daily transactions at 99.9% uptime.

AI-native engineering

Two things I have shipped matter most here. An autonomous PR review agent now reviews every pull request end-to-end - code quality, architecture, security, test coverage - leaving inline comments and verdicts with no human kickoff; it is adopted across all engineering teams and took review turnaround from hours to minutes. And a multi-agent bug-fix orchestrator - analyser, fixer, validator, and git agents behind human approval gates, each attempt isolated in its own workspace - takes a reported defect to a merged PR in under 20 minutes for well-scoped bugs.

Beyond those, 6+ production agents run across the platform (FinOps, claims processing, enterprise knowledge search, migration acceleration), cutting operational cost by 40%+ on the workloads they automate. The knowledge layer underneath is a production RAG pipeline over 50K+ enterprise documents: hybrid dense and sparse retrieval with cross-encoder reranking lifted top-5 precision from 72% to 94%, and caching plus context packing cut latency from 4.2s to under 800ms. I've fine-tuned open-source LLMs on H100 GPUs with a deployment pipeline onto GCP GPU clusters, and I'm extending that into a terminal-native coding assistant on self-hosted Gemma 3 - AI-assisted development entirely inside the enterprise security perimeter, with zero external model calls.

Before Concierto

I was Head of Products & Data Engineering at Core Compete, leading products and data engineering through the run-up to Accenture's acquisition of the company; the SaaS platform and data IP I built formed part of the acquired asset base. Earlier roles span principal engineering and architecture at YASH Technologies (10M+ daily HIPAA-compliant insurance claims), Technovert, and Barry-Wehmiller in St. Louis.

I'm an IIM Ahmedabad Senior Management Programme alumnus with a B.Tech in Computer Science. The best way to reach me is on LinkedIn.