Skip to content

SURAJ DAVARIYA — PORTFOLIO '26

Suraj Davariya

Where AI/ML platforms meet human-centered design.

A Platform Engineer, UI Visionary, A11y Advocate, AI Builder based in New Jersey.

Open Case FilesTransmit a signal

S2 · Career

Orbits

Four orbits around one core idea: complex systems, made usable.

  1. Web Development Intern

    First orbit. Learning how the web actually works — markup, style, behavior — and that craft starts with fundamentals.

    • HTML
    • CSS
    • JavaScript

  2. Frontend Engineer

    Migrating corporate banking from legacy stacks to a React + Redux SPA — and debugging it in production with real user sessions.

    • React
    • Redux
    • YUI
    • TeaLeaf
    • NVDA
    • JAWS

  3. UI Lead

    Leading UI teams on federal health platforms — Angular at scale, mentoring juniors, and making ADA compliance a build habit, not a gate.

    • Angular
    • RxJS
    • TypeScript
    • NVDA
    • WAVE
    • SonarQube

  4. Software Engineer II

    AI/ML platforms for Markets Operations — productionizing data-science models and rendering millions of rows in seconds, on AWS end to end.

    • AWS
    • Kafka
    • SageMaker
    • Airflow
    • Terraform
    • Python

MS-CS · NYIT 2016 — AWS Certified Developer & Cloud Practitioner — IEEE Member

S3 · Case Files

Data Streams

Four problems, abstracted to what matters: the constraint, the approach, the result. The specifics stay behind the firewall.

Decorative animation: a stream of particles flowing into the silhouette of a data table, row by row.

CF-01 · JPMORGAN CHASE

Millions of rows. Seconds to render.

Context
Markets operations teams needed to explore live datasets that had outgrown every off-the-shelf grid.
Role
Engineer on a greenfield data platform — frontend and pipeline.
Approach
Streaming end to end: Kafka in, AWS Flink and Deephaven transforming, a virtualized React grid out.
Impact
Millions of rows render in seconds, straight from the stream.

1M+rows, rendered in seconds

  • React
  • Deephaven
  • Kafka
  • AWS Flink

Decorative animation: an MLOps pipeline of connected AWS service nodes with pulses flowing from SageMaker through Glue and Lambda to EKS.

CF-02 · JPMORGAN CHASE

From notebook to production.

Context
Data-science models were proving value in notebooks but had no reliable road to production.
Role
Platform engineer productionizing ML delivery for Markets Operations.
Approach
SageMaker for models, Airflow DAGs for orchestration, Lambda + S3 events feeding Kafka — all provisioned as Terraform code.
Impact
Models ship as an engineering pipeline, not a handoff.

100%infrastructure as code

  • SageMaker
  • Airflow
  • Terraform
  • Python

Live demo: three switches that apply reduced motion, high contrast, and keyboard mode to this entire site.

CF-03 · FDA

WCAG 2.2 AA

Immersive, and accessible to everyone.

Context
Federal health platforms are required — legally and morally — to work for everyone.
Role
UI lead owning ADA compliance and mentoring the team.
Approach
Built and audited with NVDA and WAVE in the loop from day one, not as a final gate.
Impact
Accessibility became a build-time habit — the same discipline this site runs on. Try it live:
  • Angular
  • RxJS
  • NVDA
  • WAVE

Decorative animation: a component tree assembling node by node while state pulses outward from a central store.

CF-04 · WELLS FARGO

React + Redux SPA

Banking-grade React at scale.

Context
A corporate-banking workflow trapped in a legacy YUI + Handlebars stack.
Role
Frontend engineer on the SPA migration team.
Approach
Rebuilt as a React + Redux single-page app, using TeaLeaf to replay and debug real production sessions.
Impact
A maintainable component system in place of page-level legacy code.
  • React
  • Redux
  • TeaLeaf

S4 · Labs

Open Laboratory

Public experiments, pulled straight from GitHub — and the HELIOS core itself, with the controls handed over to you.

Live core controls

// the sun behind this page is listening

100%
×1.0
SOLAR

S5 · Arsenal

Constellation

Ten years of tools, mapped. Focus a star to light up what it connects to.

Frontend

Cloud & AWS

AI-ML Platform

Quality & A11y

S6 · Insights

Signals

Field notes on accessible immersion and AI/ML platform engineering — written from the build, not the sidelines.

All signals →

S7 · Transmission

Open a channel

New problems, collaborations, or a conversation about accessible immersion — the uplink is listening.

ENGINEERING INTELLIGENCE

© 2026 Suraj Davariya · engineered with Next.js + R3F · fully keyboard-navigable