About Devansh Singh

Devansh Singh

Bengaluru, India • Full-Stack Developer & AI Engineer

Work-Integrated Learning B.Tech at Kalvium • Microsoft Azure Developer Community Secretary • Former Google Student Ambassador

Who I Am

My relationship with software began with curiosity about how complex digital systems organize chaos. Early on, I was drawn to the mechanics of networked communication: how two isolated machines on opposite ends of the world can exchange cryptographic signatures in milliseconds, or how streaming data feeds can coordinate algorithmic decisions without human latency.

Today, I am based in Bengaluru—India's technological nerve center. I am pursuing an industry-integrated Bachelor of Technology in Computer Science through Kalvium, where software engineering isn't treated as an academic abstraction, but as a daily craft of building real products from day one.

How I Approach Software Engineering

I am skeptical of engineering choices driven purely by industry fashion. In an era where web apps routinely ship multi-megabyte JavaScript bundles for simple text displays, I believe strongly in restraint, semantic clarity, and predictable execution:

• Server First, Client When Justified: Content should exist in the initial HTML response. Search engine crawlers, assistive screen readers, and users on low-bandwidth mobile connections should never have to wait for client-side JavaScript hydration simply to read text.
• Observability Over Magic: As we delegate more cognitive tasks to AI agents and autonomous LLM loops, opacity becomes an existential bug. Systems must expose verified telemetry—every tool call, network handshake, and decision branch should be auditable.
• Performance as an Ethical Consideration: Fast page loads and minimal battery consumption respect the user's device and time.
• End-to-End Ownership: Whether it's configuring Docker containers, optimizing SQL/NoSQL schemas, tuning WebSocket backpressure, or refining typographic line-heights, great software comes from understanding the whole stack.

What I Build & Research

My engineering work spans three complementary dimensions:

1. Distributed Full-Stack Backends: Scalable applications built on Node.js, Next.js, Express, Redis Pub/Sub, MongoDB, and WebSockets. I have built secure peer-to-peer WebRTC messengers (Chatty) and financial signal ingestion pipelines (StockBeacon).
2. Agentic LLM Tooling: Developing background daemons and Model Context Protocol (MCP) servers (AI Bot, MCP AI Server Agent) that allow frontier models to safely navigate local developer workspaces, execute vetted shell commands, and automate system routines.
3. Graph Representation Learning: Researching the use of Graph Neural Networks (GNNs) and attention mechanisms to counter transactional fraud and identify laundering chains across massive financial payment topologies.

Community Leadership & Sharing Knowledge

I believe knowledge compounds when shared openly. As Secretary of the Azure Developer Community (Alliance University), I have organized and led hands-on technical workshops, Azure bootcamps, and hackathons reaching over 300 students. As a former Google Student Ambassador, I guided peers through cloud and machine learning certifications.

In 2025, I served as the Lead Technical Coordinator for a global ML hackathon with over 500 participants, managing API credits, GPU cluster allocations, and technical judging criteria.

Beyond the Terminal: Couplets, Books & Philosophy

I am not just someone who writes code. Human beings are multi-dimensional, and my creative life is enriched by interests far removed from software engineering:

• Urdu & Hindi Poetry (Shayari): I write original Ghazals and couplets. Poetry demands the same precision as systems programming: every word must carry weight, rhythm, and purpose.
• Reading: I enjoy reading about tech ethics, systems thinking, history of computing, and philosophy of mind.
• Current Learning: I am currently diving deeper into systems programming in Rust, exploring formal verification for tool-calling agents, and refining high-density data visualization techniques.