I’ve been thinking a lot lately about how I got here - building AI automation systems, optimizing cloud costs, working with technologies I didn’t even know existed four years ago. The path from fresh graduate to where I am now wasn’t linear, and honestly, it wasn’t what I planned either. But looking back, every detour made sense.
The Beginning: Choosing Between Two Paths
When I graduated from MIT Pune in 2020, I had two job offers: Software Engineer at Cybage and System Engineer at IBM. I spent days making pros and cons lists, talking to friends, trying to figure out the “right” choice. IBM had the brand name, but Cybage’s role felt clearer - I could actually see myself writing code and building things there. So I chose Cybage, thinking my career path was set.
Then COVID hit, and everything stopped.
Just as I was about to start, I got an email from Cybage saying onboarding would be “delayed until further notice.” No timeline, no guarantees. I was 22, freshly graduated, and suddenly unemployed before I even started. The carefully planned career trajectory I’d imagined? On hold indefinitely.
The Accidental Web Developer
During those uncertain months of lockdown, I needed something to do. My dad’s company needed a website, so I started building one for them.
Here’s the thing though - web development was never my dream.
I was the kid who loved data structures, algorithms, and system-level programming. I wanted to work with data analytics and automation, not make things look pretty on a screen. But I had time and my dad needed help, so I dove in.
I spent those months learning HTML, CSS, JavaScript - the basics. Nothing fancy, just functional websites. I was Googling everything, breaking things constantly, and slowly figuring it out.
Sometimes the path you resist is exactly the one you need to take.
Looking back, those months taught me something important - I just didn’t know it yet.
December 2020: Finally Starting
Eight months after I was supposed to start, I finally joined Cybage in December. The onboarding was completely virtual - all of us new hires sitting in our respective homes, cameras on, learning to be developers through a screen. They trained us as full-stack developers: Angular for frontend, Node.js for backend, database design, API architecture, the whole stack.
There were 40 people in my batch. Some had internships at big companies, some had impressive GitHub profiles, some had won hackathons. And there was me - the guy who’d spent lockdown building basic websites for his dad. I remember thinking I was probably middle of the pack, maybe lower. So I worked harder than I’d ever worked, not because I wanted to be the best, but because I was terrified of being the worst.
After weeks of training, projects, and assessments, they published the scores.
I came first.
Out of 40 people. I stared at that screen for a solid minute, convinced there was a mistake. There wasn’t. A few days later, I got called into a meeting - they had an opportunity to work on Google projects. Of course I said yes.
Two Years at Google (Through Cybage)
For the next two-plus years, I worked on Google’s marketing properties. The work was mainly frontend - content-heavy, visual-heavy, lots of UI polish. We built marketing pages, interactive experiences, promotional campaigns. I got really good at Angular, then learned React on my own, and something clicked. The component model, the ecosystem, the way everything flowed - it just made sense. Web development wasn’t just tolerable anymore, I actually liked it.
But here’s the thing about comfort - it can be a trap. I was learning, yes. Getting better, absolutely. My team was happy, the Google stakeholders gave positive feedback, and I’d found my groove. But that voice in the back of my head, the one that loved data automation and wanted to build complete systems, kept getting louder. My training was in full-stack, but my daily work was almost entirely frontend. The backend skills, the database design, the system architecture - all sitting there, unused.
I started doing side projects in my free time. Building APIs, setting up databases, writing automation scripts. Not for work, not for money, just because I wanted to. That’s when I realized I didn’t just want to build interfaces - I wanted to build entire systems. Frontend, backend, database, deployment, optimization, everything. And with AI and machine learning exploding everywhere, I couldn’t shake the feeling that I was building marketing pages while the future was being built somewhere else.
The Gamble: Leaving for Grad School
I’d had this dream since college: travel, study abroad, dive deeper into computer science. Not just learn frameworks and libraries, but really understand the fundamentals - theory, research, the kind of depth you don’t get from tutorials. A Master’s degree abroad - US, Canada, or Australia - that was the goal.
My parents were my backbone in this decision. They supported me throughout, even though it meant I’d be leaving a good job and going halfway across the world. Their support made the leap possible.
I applied to universities and waited. In March 2023, I got my first admit letter from UT Arlington. More admits came over the next few weeks from other universities I’d applied to. I had to make a decision, and I also had to figure out: Master’s or PhD? I chose Master’s first, figuring I’d keep the PhD door open but not commit yet.
But I said yes anyway. My teammates were supportive but confused - why would anyone leave a stable job to be a broke student again, in a new country no less? Even I thought I might be making a mistake. But comfort is the enemy of growth, and I was too comfortable. So in August 2023, I quit my job, packed my bags, and flew to Texas.
UT Arlington: Back to Being a Student
Being a student after working professionally for two years is weird. You go from having a salary, your own money, independence, and respect as a working professional to being broke, living with roommates, eating ramen, and being treated like you don’t know anything. The first few weeks were rough - culture shock, homesickness, financial stress, imposter syndrome hitting harder than ever.
My classmates were brilliant. Some had published papers, some had worked at FAANG companies, some were straight out of undergrad but knew more about machine learning than I did. I questioned my decision daily. But then classes started, and something shifted.
I specialized in AI and Intelligent Systems - finally learning what I’d always wanted to learn. Machine learning algorithms, neural networks, natural language processing, computer vision. But I didn’t stop there. I took database systems to understand data at scale, distributed systems to learn infrastructure that doesn’t break, software engineering to write maintainable code for real products. 36 credits, countless late nights, so much coffee.
And projects - so many projects. Not just classroom assignments but real things. Full-stack applications with authentication, API design, database optimization, deployment pipelines. Backend systems with REST APIs, GraphQL servers, message queues, caching layers. LLM integrations that actually worked in production, not just demos. End-to-end systems with everything working together.
This is when everything I’d learned started to merge. The frontend skills from Cybage, the full-stack training, the backend interests I’d always had, the data automation dreams from college, and now AI and machine learning. I wasn’t just a frontend developer anymore, not just a backend developer either. I was becoming someone who could build complete systems from the ground up.
The Job Search Reality
As graduation approached, I started applying for jobs, and reality hit hard. The 2024-2025 tech job market was brutal - layoffs everywhere, hiring freezes, “entry-level positions requiring 3+ years experience.” I sent out 200+ applications, maybe more. I lost count. Rejections, automated emails, ghosting, “we’ve decided to move forward with other candidates.”
There were days I wondered if I’d made a terrible mistake. I’d left a secure job at Cybage to be unemployed in a foreign country with student debt. Brilliant move. But then I got a response from Smart Rewards.
Smart Rewards: Building Systems That Matter
The interview was different. They didn’t just want to know if I could code - they wanted to know if I could solve real business problems. “We need automation. Our manual processes are killing us. Can you build systems that actually save us time and money?” Could I? This was literally what I’d been training for.
From day one at Smart Rewards, I was doing exactly what I’d dreamed about. Building automation workflows with n8n, connecting different tools, making processes that used to take hours happen in seconds. Creating FastAPI backends - fast, type-safe Python APIs that power our automation systems. Designing systems for multiple users with proper authentication, authorization, logging, monitoring, error handling. Integrating AI where it actually makes sense, not just throwing GPT at everything.
And then came the project that proved everything I’d learned was worth it. We needed to process videos - 360 videos per month. The existing solution was costing us $700 per month using AWS Lambda and S3. My manager asked if I could make it cheaper. I spent weeks analyzing the workflow, understanding bottlenecks, researching alternatives. This wasn’t just a coding problem - it was system design, cost optimization, a business problem.
I rebuilt the entire pipeline. Changed the architecture, optimized the processing, made different technology choices.
The result? $10 per month.
From $700 to $10. A 98% cost reduction, $8,280 saved per year, for one system. Same 360 videos per month, same quality, same speed, better reliability. That moment when I showed my manager the first month’s bill and his eyes went wide - “Wait, is this for real?” - that was validation.
All the years learning frontend, the backend self-study, the gamble of leaving Cybage, the stress of grad school, the brutal job search - all of it led to this. Building systems that actually matter.
The Hard Truth Nobody Tells You
Here’s something I wish someone had told me when I started at Cybage, fresh out of college and excited about my first job.
You’re replaceable.
I know that sounds harsh, but hear me out. You’re just one grain of sand in a vast desert. There are thousands, maybe millions of developers like you. Many are smarter. Many have better credentials. Many have more experience. That’s just reality.
But here’s the thing - this truth isn’t meant to be depressing. It’s meant to be liberating.
If you’re just one grain of sand, you have nothing to lose and everything to gain. You can take risks. You can bet on yourself. You can leave comfortable jobs for uncertain futures. Sometimes you’ll get opportunities fast, sometimes you’ll have to wait for the right moment, and sometimes you’ll send 200 applications and get three responses.
None of that matters. What matters is what you do with the time you have.
The only constant: keep learning, never stop competing. Not against others, but against who you were yesterday.
When I joined Cybage, I didn’t know React. I learned it. When I wanted to go deeper into backend, I taught myself outside work hours. When I decided I needed to understand AI, I went to grad school. When I needed to prove I could solve real problems, I reduced a $700/month cost to $10/month.
Every step was uncomfortable. Every transition was scary. Every decision felt like a gamble.
But that’s the point. Growth doesn’t happen in comfort zones.
What I’m Building Now
I wake up excited about work, and that’s not something I could say every day at Cybage, even though it was a good job. At Smart Rewards, I’m building systems that matter - automation that saves real money, not theoretical optimizations but actual dollars saved and hours returned. AI that runs in production, not demos or proofs-of-concept but real systems handling real workloads. Infrastructure designed to scale with clean architecture, maintainable code, and proper monitoring. Tools that empower multiple people, not just scripts on my laptop but shared systems with authentication, permissions, documentation, support.
Every system I build has constraints: it has to work for multiple users, be maintainable when I’m not around, be cost-effective, and actually solve the problem. There’s no room for flashy tech that doesn’t deliver, no space for over-engineering. Just good, solid systems that do what they’re supposed to do. And I love it.
I’m also working on side projects - building my own tools, exploring new technologies, writing these articles to document what I’m learning. Current projects include a custom Django CMS because no existing CMS solves my specific needs, migrating systems from Firebase to PostgreSQL to learn database migration at scale, component libraries and design systems because I still love good frontend work, and LLM-powered automation experiments because AI is getting wild.
The Journey Continues
If you’d told 2020-me, the guy building basic websites during COVID lockdown, that in 2026 I’d be building AI automation systems that save tens of thousands of dollars per year, I don’t think I would have believed you. The path wasn’t what I expected. I thought I’d be a data analyst but became a frontend developer. I thought I’d just do Angular but fell in love with React. I thought I’d stay in frontend but dove deep into backend. I thought I’d work at Cybage for years but left after two. I thought I’d go to a top-tier university but went to UT Arlington and learned just as much. I thought I’d struggle to find a job after graduation but found one where I’m doing exactly what I love.
The lesson? The path you start on isn’t always the path you end up taking, and that’s not just okay - that’s how it should be. I started with frontend because that’s what the job needed, got good at it because I don’t do things halfway, then went deeper into backend because that’s what I was curious about, then learned AI because that’s where the industry was going. Each step built on the previous one, each skill added to the toolkit, each experience informed the next decision.
Now I’m a full-stack engineer who specializes in AI-powered automation. That’s a title I couldn’t have imagined in 2020, but it’s exactly where all the pieces fit together.
For Anyone Reading This
If you’re early in your career, thinking about switching paths, or wondering if you should take that leap, here’s what I’ve learned: Your first job won’t define your career - I started building marketing pages and now build cost-optimization systems and AI automation. Sometimes you have to leave comfort - leaving Cybage was scary, leaving India was scarier, being a broke student after earning a salary was the scariest, but none of my growth would have happened in my comfort zone.
Keep learning, always. The tech stack I use today - FastAPI, n8n, modern LLMs, cloud infrastructure - literally none of it existed or I didn’t know it existed when I started at Cybage in 2020. Technology moves fast, and you have to move with it. Focus on impact over titles - I could be a Senior Software Engineer still building nice UIs, but instead I’m building systems that could save $8,400 per year. Which sounds more impressive in an interview? Which actually matters more?
Take the gamble when it feels right. Not every gamble pays off, but the gambles you don’t take are guaranteed to return nothing. I gambled on grad school, and it paid off. You won’t know until you try. And document your journey - I wish I’d started writing these articles earlier. Documenting what you’re learning, building, and struggling with is valuable for you and for others following similar paths.
Looking Forward
This story isn’t over. I’m 26 with decades of building ahead of me. There are still technologies I want to learn, systems I want to build, problems I want to solve, mistakes I’ll make, and lessons I’ll learn the hard way. Maybe I’ll do a PhD eventually, maybe I’ll start my own company, maybe I’ll specialize even further in AI infrastructure, maybe I’ll go in a completely different direction. I don’t know, and that’s exciting.
What I do know is this: I’m no longer the confused college grad trying to choose between two job offers. I’m not the frontend developer wondering if there’s more to tech than pretty UIs. I’m not the grad student questioning if leaving a good job was a mistake. I’m someone who builds systems that matter, someone who takes complex problems and finds elegant solutions, someone who’s still learning, still growing, still pushing.
The journey from Google marketing projects to AI automation wasn’t linear. It was messy, uncertain, sometimes scary.
But I wouldn’t change a thing.
This is the first article in a series where I’ll share what I’m building, what I’m learning, and the lessons from the trenches of production systems and AI automation.