This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI
About Me
My work is driven by a simple belief: the most memorable experiences are the ones that feel personal. I’m Vishwajeet Singh Thakur, an entrepreneur and software engineer who views technology not just as a tool for scale, but as a canvas for meaning.
In a world where digital interactions often feel generic and impersonal, I believe we’re standing at a turning point. Modern technology is powerful enough to recreate the nuance, warmth, and relevance of human connection - if used intentionally. That conviction led me to trade a stable career in 2025 for a singular mission: embedding personalization into the DNA of digital experiences.
This philosophy shapes everything I build, from my venture Invysia to the portfolio you’re exploring here. Each project reflects a focus on precision, empathy, and thoughtful design - demonstrating that deeply personal, engaging digital experiences are no longer slow, expensive, or experimental. With the right technology and mindset, they can be built with a speed and quality that was previously unimaginable.
This portfolio is not just a showcase of my skills but is a statement of intent.
Portfolio
This portfolio is not a static website; it’s a living, adaptive experience. What you see here is a curated expression of the same body of work, reshaped dynamically based on who you are and how you choose to engage with it. Visitors can select from four distinct personas and sixteen languages, resulting in sixty-four unique variations of the portfolio, each tailored to a different context, mindset, and intent.
 |
 |
 |
Behind the scenes, the site is powered by an agentic backend that generates content on demand at the moment of interaction. Using a lightweight, high-performance generative model (Gemini 2.5 Flash), the system balances speed and cost efficiency while delivering coherent, relevant narratives in real time—ensuring that no two visits feel exactly the same.
The intelligence goes deeper than presentation. Its multi-agent architecture syncs with my personal GitHub repository through an MCP server, automatically discovering, summarizing, and contextualizing projects as they evolve. This means the portfolio stays current by design, reflecting real work as it changes, rather than snapshots frozen in time.
The interface is also attentive to user behavior. It adapts to how you navigate - anticipating intent, reducing friction, and offering shortcuts when needed. If you skim quickly, it proposes a concise overview. If your question isn’t answered through predefined paths, you can engage directly with the agents through conversation.
All of this is wrapped in a deliberately minimal, clutter-free interface. The goal is restraint, not absence, removing visual noise to make space for meaningful interaction. The result is a portfolio that doesn’t just show what I’ve built, but actively demonstrates how I think about technology, experience, and personalization.
Explore the live portfolio experience:
https://portfolio-frontend-268314723675.us-central1.run.app/
How I Built It
This project is a reflection of how my development workflow has evolved alongside modern AI tooling. Before 2026, my setup revolved around a familiar combination of IntelliJ IDEA or VS Code paired with Cline and Claude Opus. Over time, however, I found myself optimizing not just for code quality, but for momentum—and that led me to rethink the role of the IDE itself.
In late 2025, I transitioned fully to Antigravity IDE, which has since replaced both IntelliJ and VS Code as my primary environment. It allows me to work closer to intent than syntax, reducing friction between thought and execution. For implementation-heavy tasks, I rely on Claude Opus 4.5 for its consistency and depth in reasoning. For exploratory and research-oriented work, Gemini Pro has become my model of choice. I also use the Gemini CLI directly within the terminal, enabling fast, contextual search without breaking focus or switching contexts.
From a technical standpoint, the stack behind this portfolio is intentionally lean. The front end is built with Next.js and TypeScript, providing a performant, type-safe foundation for dynamic UI. The agentic backend is implemented in Python using Google’s Agent Development Kit (ADK), which orchestrates content generation, behavioral adaptation, and real-time data integration.
The result is a system where tooling, models, and architecture work in concert - each chosen not for novelty, but for how effectively they support rapid, high-quality iteration.
For those curious about how it works, the code lives here:
https://github.com/vishwajeet22/portfolio
What I'm Most Proud Of
What stands out most to me is not any single feature, but the compression of effort into outcome. Using this toolchain, I was able to design and build something that is meaningful, interactive, and expressive of my values, vision, and mission - end to end - in just two days.
Not long ago, a project of this scope would have taken weeks of fragmented work. This build is proof of how far the craft has advanced: when the right tools are paired with a clear intent, speed no longer comes at the expense of depth or quality.
Top comments (0)