Dewansh Singh Thakur
Software Engineer
I started as a full-stack engineer, but over the last few years I've increasingly moved toward the systems underneath the product — infrastructure, distributed systems, AI tooling, networking, and performance. Currently a Software Development Engineer II at Memorang, working on AI-powered education products and the infrastructure behind them.
I believe in understanding fundamentals before abstractions, measuring before optimizing, and building systems that last.
Engineer at Core
How I approach software development
Full-Stack Engineer
Comfortable across the entire stack — from React frontends to Node.js backends, databases, and deployment pipelines. I ship complete features, not just fragments.
System Thinker
I don't just write code; I design systems. Understanding trade-offs, data flow, and bottlenecks before implementation leads to better architecture.
Performance Focused
Fast applications aren't accidental. I profile, measure, and optimize — whether it's rendering millions of items in React or reducing backend latency.
Architecture Oriented
Scalable systems require thoughtful design. I think about modularity, separation of concerns, and making code that's maintainable at scale.
What I Build
Areas where I create impact
AI-Powered Products
Features and workflows around LLMs, agents, and AI-assisted interfaces — from education tools to developer tooling.
Backend Systems
APIs, commerce platforms, authorization systems, and databases that power multi-tenant SaaS products.
Infrastructure & Tooling
Self-hosted services, containerized deployments, and the developer tooling that makes systems easier to operate.
Distributed Systems
Real-time sync, message passing, and networked systems — built to handle failures and stay available.
Performance-Optimized Software
Rendering efficiency, low-latency backends, and systems where every millisecond matters.
Home Lab Experiments
Proxmox virtualization, e-ink dashboards, and hands-on infrastructure to understand systems from the hardware up.
How I Think About Systems
Principles that guide my engineering decisions
- →Simplicity over abstraction — The best code is code you don't have to explain.
- →Measure before optimizing — Assumptions about performance are often wrong. Profile first.
- →Understand internals before adopting tools — Frameworks are great, but knowing what's under the hood prevents surprises.
- →Security is non-optional — Every system should be built with security in mind from day one.
- →Code is read more than written — Clarity and maintainability matter more than cleverness.
- →Test the edges, not just the happy path — Real systems break in unexpected ways.
Technical Stack
Focused tools I work with
Languages
- TypeScript
- Python
- Rust
- JavaScript
- Dart
Systems & Infrastructure
- Linux
- Docker
- Kubernetes
- PostgreSQL
- AWS / GCP
- Proxmox / Nginx / Tailscale
Application
- Next.js
- React
- Node.js
- Django
- Flutter
AI
- LangGraph
- LangChain
- MCP
- LLM APIs
Beyond Code
What drives curiosity
Engineering Curiosity
Always exploring how things work — from networking protocols to hardware internals. I learn by breaking and rebuilding.
Home Lab Experiments
Running a Proxmox homelab with self-hosted services, containerized workloads, and hardware projects like the distributed e-ink dashboard.
Photography
Capturing moments and experimenting with composition. It's another form of creating something from nothing.
Coffee Obsession
Good coffee, good code. There's something about the ritual that makes complex problems feel approachable.
Mechanical Watches
Appreciating precision engineering and craftsmanship. The intersection of mechanical complexity and timeless design resonates with how I think about software.
Long-Term Direction
Where I'm heading
I'm moving toward deeper work in distributed systems and performance engineering. Building software that scales across machines, handles failures gracefully, and stays fast under load is the kind of challenge I want to tackle.
The intersection of AI and infrastructure fascinates me — not just using AI tools, but understanding how to architect systems that support AI workloads efficiently.
I'm also drawn to autonomous systems — software that makes intelligent decisions without constant human intervention. The elegance of well-designed automation is something I want to explore further.
Long-term, I see myself contributing to systems that operate at scale, where every millisecond matters and reliability is critical. That might mean distributed databases, real-time infrastructure, or mission-critical applications in specialized domains.