How I Learned to Build Fast Software and Write Books

From my first programming lessons in West Bengal to saving 12 million dollars on enterprise cloud server bills, studying 50 startups in India, and building in-memory engines.

Early Curiosity and University Studies

My interest in technology started with a simple question: why do computers make people wait? As a teenager, watching spinning loading icons on slow desktop computers made me wonder why machines capable of calculating millions of operations per second would freeze over basic tasks. When I entered college to study computer applications at West Bengal University of Technology (WBUT), I spent nights reading how computer chips, memory circuits, and operating systems communicate at the hardware level.

I realized early on that software gets slow not because hardware lacks power, but because software developers layer unnecessary abstractions, make redundant disk calls, and pull in bloated libraries. In 2016, I completed my Bachelor of Computer Applications degree and graduated as the Rank 1 Valedictorian across the entire university batch. That honor taught me that disciplined focus, clear thinking, and deep curiosity can open doors to solving hard engineering problems.

Enterprise Apprenticeship: Learning Where Technology Breaks

After graduating, I chose to work within large global enterprises to witness firsthand how mission-critical software behaves under heavy real-world scale and where corporate engineering departments fail:

  • Deloitte (2016 - 2018): As a Business Technology Analyst, I helped organize financial reporting automation for more than 30 enterprise clients. I saw how manual Excel workflows cost multinational corporations millions of dollars in administrative overhead and error correction. By writing streamlined automated scripts, our team reduced recurring reporting effort by 20%.
  • Procentris (2018 - 2020): Serving as Technical Lead, I led an engineering squad connecting clinical records across 57 healthcare agencies. Dealing with patient health data taught me that reliability is not an abstract metric; people's wellbeing depends on system uptime. We optimized relational database schemas and improved health records dispatch speed by 30%.
  • Clean Harbors (2020 - 2022): As Data Product Owner, I managed operational systems supporting 500 environmental service operators. When the pandemic hit and demand surged, our team maintained continuous 99.5% operational uptime by re-architecting connection pooling and circuit-breaker patterns, earning me the Employee of the Month award.
  • PayPal & eClerx (2022 - 2023): Leading fintech automation pipelines, I architected automated reconciliation systems for consumer payment portfolios. We eliminated more than 10,000 hours of manual reconciliation per year, replacing tedious spreadsheet comparisons with real-time programmatic verification.

The Twelve Million Dollar Cloud Lesson

At CAI, I was tasked with conducting a deep audit of enterprise cloud computer spending across multi-region AWS and Azure accounts. What I discovered was startling: hundreds of expensive server instances were running 24 hours a day, 7 days a week, with nobody using them. Staging servers were left spinning over weekends, test databases were provisioned with 64 CPU cores when 2 cores were sufficient, and unattached elastic block storage volumes were quietly billing thousands of dollars every month.

Our team implemented automated shutdown scripts for non-production environments, rightsized production compute tiers to match real p95 traffic, and renegotiated reserved instance commitments. Over 18 months, our audit eliminated twelve million dollars ($12,000,000) in annual cloud waste. That experience solidified my core architectural philosophy: good software is lean, fast, respectful of hardware, and capital-efficient.

Academic Research at Liverpool Business School

To understand the intersection of technology architecture and company survival, I pursued a Master of Business Administration (MBA) at Liverpool Business School (LJMU) in the United Kingdom. For my master's dissertation, I conducted empirical research analyzing 50 real-world Indian startups across deeptech, SaaS, and e-commerce.

The research findings were conclusive: startups that relied heavily on continuous venture debt and burned cash on bloated cloud infrastructure experienced high failure rates. In contrast, sovereign founders who maintained lean small teams, prioritized gross margins, and built capital-efficient architectures survived market downturns and built enduring, profitable businesses.

Building Fraoula and Writing Canonical Guides Today

Today, as Co-Founder and Chief Product Officer at Fraoula in Bangalore, I lead the architecture of our proprietary in-memory data processing platform. Our engine checks incoming marketing and transaction data in under 350 milliseconds by holding active working sets in memory ring buffers rather than hitting slow disk databases.

Alongside building software, I write comprehensive canonical essays and published books. I believe in stripping away industry buzzwords and explaining technology in plain 5th-grade English. Whether through my published volumes, our weekly newsletter, or private advisory calls, my mission is to help the next generation of engineers build sovereign, fast, and honest software.

Helpful Pages to Read Next

← Return to Home Contact Arkajit Das →