SENIOR SOFTWARE ENGINEER · LEGALZOOM

I build product systems that make complexity feel simple.

I’m Debasis Nath, a full-stack engineer with deep frontend-platform roots. I work across Go, React, and TypeScript—and I’m extending that systems mindset into agentic AI and MCP.

Explore selected work GitHub
15+products supported
80%faster integration
95%faster staging deploys
60%less manual intervention

SELECTED SYSTEMS / 01—03

Work measured by what changed.

Sanitized stories from production work—focused on constraints, engineering decisions, and outcomes instead of feature lists.

01PLATFORM ENGINEERING

A shared UI platform for a 15+ product fintech suite.

Built common components, observability foundations, and configuration-driven localization that helped independent teams ship with a more consistent system.

  • React + TypeScript
  • 15+ products
  • 3× localization cycles
02PRODUCT ARCHITECTURE

Decoupling reconciliation workflows from a monolith.

Led a headless-hook architecture with React and Zustand, turning a tightly coupled product surface into composable modules that teams could integrate independently.

  • Headless architecture
  • 80% faster
  • 2+ weeks saved
03WORKFLOW AUTOMATION

Making enterprise tax filing more reliable.

Shipped a TypeScript microfrontend and hardened Cypress automation around complex government-site workflows, improving completion while reducing human intervention.

  • Microfrontend
  • +77.9% success
  • −60% intervention

AI SYSTEMS LAB / BUILDING IN PUBLIC

From product engineer to production AI systems engineer.

A build-first journey through evaluation, agents, production ML, inference, and Edge AI. Progress is earned through tested artifacts, benchmarks, and engineering reports—not course completion badges.

Follow the journey
CURRENT PHASE · FOUNDATIONS + GEN AI SYSTEMS
IN PROGRESS

AI Systems Engineering Journey

A public roadmap and engineering journal spanning evaluated RAG, bounded agents, production ML, CUDA, TensorRT, and device-aware inference.

A01
BUILT

Nasdaq Learning Lab

An offline-first learning product with sequential modules, quizzes, calculators, journaling, data portability, and installable PWA behavior.

A02 ↗
NEXT EVIDENCE

Evaluated RAG system · constrained tool agent · benchmark report

EXPERIENCE / 2018—NOW

A platform mindset, across the stack.

01

Senior Software Engineer

LegalZoom

Building full-stack product systems across Go, React, and TypeScript while expanding into AI-enabled workflows.

02

Software Engineer II · UI

Clear

Owned frontend platform and product architecture across shared libraries, reconciliation, enterprise filing, localization, observability, and delivery tooling.

03

Software Engineer I

Clear

Built fintech product surfaces, notification and messaging infrastructure, and reusable React systems across the product suite.

04

Software Engineer

MountBlue Technologies

Developed production web applications and established the full-stack foundation behind a platform-focused career.

OPERATING PRINCIPLES

Build the system. Measure the change. Explain the tradeoff.

01

Evidence over theatre.

A benchmark, shipped artifact, or measured outcome beats a long technology list.

02

Platforms serve teams.

Good architecture reduces the cost of change for the engineers and products around it.

03

Reliability is a feature.

Observability, fallbacks, evaluation, and failure handling belong in the design from day one.

EDU / ACTIVE

BS in Data Science

Indian Institute of Technology Madras · Ongoing

Formal data-science study alongside independent production AI and Edge AI engineering.

LET’S BUILD THE NEXT SYSTEM

Have a hard product or platform problem?

Start a conversation github.com/devoficial