M.S. CS @ Texas A&M | ex Walmart Global Tech

Building calm, scalable systems with a craftsperson's eye.

I design software that feels effortless under pressure. My sweet spot is the moment where infrastructure, performance, and product clarity meet. I iterate deeply, measure carefully, and make tools feel human.

See the work System map
Based in College Station, TX
Focus Platform, performance, cloud
Curiosity NLP + human-centered AI
Alisha Raj
Intro

Hi, I'm Alisha. I build reliable systems where performance, clarity, and human-centered tooling work together.

Currently

Exploring cloud systems, NLP workflows, and the craft of making complex software feel simple.

Open to bold, high-impact problems.

Signature work, system tradeoffs

I care about the throughput, constraints, and failure modes that make systems reliable.

GraphQL CRM at 20k+ RPS

Moved from N+1 REST to typed GraphQL, stabilized schema contracts, and reduced cache poisoning while keeping 20k+ RPS steady.

C# · EF · GraphQL · schema governance · cache hygiene
40% payload reduction 20k+ RPS Schema hardening
Explain like:

Choose a lens to rewrite this project.

Pricing performance at Walmart

Rebalanced RU usage, introduced batching and partition tuning, and removed hotspots to lift ingestion throughput 4x.

Java · CosmosDB · Kafka · RU tuning · partition strategy
4x ingestion speed 23% fewer DB calls Lower throttling
Explain like:

Choose a lens to rewrite this project.

Club recommender for 1,200+ orgs

Combined BM25 with SBERT, added a classification layer, and shipped a full-stack pipeline for 1,200+ orgs.

React · FastAPI · NLP · hybrid retrieval
1,200+ orgs 80% faster discovery Hybrid BM25 + SBERT
Explain like:

Choose a lens to rewrite this project.

How I build systems

My portfolio reads like a narrative of decisions, not just a list of features.

Scale is a choreography

I map bottlenecks early, add buffers, and tune capacity so traffic spikes become predictable instead of chaotic.

Reliability is a feedback loop

Canary deployments, alerts, and runbooks are part of the feature. No surprises, only signals.

Data is a contract

Typed schemas and clean APIs keep teams aligned and reduce time-to-clarity for every stakeholder.

UX is an engineering outcome

Performance budgets and careful defaults give the interface a calm, intentional feel.

System map

A living constellation of the tech I use to shape reliable systems.

Currently highlighted

Kubernetes

Where I shape deployments, traffic rules, and resilience strategies.

scale reliability automation
Tap a node to explore a new tool or platform.

Case studies, not projects

Each story includes constraints, tradeoffs, and what broke along the way.

Pricing throughput revival

  • Problem: ingestion lag during peak loads.
  • Constraints: fixed RU budget, strict latency SLAs.
  • Tradeoffs: batching writes vs. freshness.
  • Metrics: 4x ingestion, 23% fewer DB calls.
  • What broke: partition hotspots; solved with redistribution.
  • Learning: measure p95 before tuning.

CRM API consolidation

  • Problem: endpoint sprawl and inconsistent payloads.
  • Constraints: 20k+ RPS, zero downtime migration.
  • Tradeoffs: schema rigor vs. rapid iteration.
  • Metrics: 40% payload reduction.
  • What broke: downstream caching assumptions.
  • Learning: ship observability with schema changes.

What if traffic spikes 100x?

A small simulator to show how I think about load, cache, and cost.

1500
RPS
140ms
p95 latency
88%
cache hit rate
$1.2k
daily infra cost

Bottleneck shift

At this load, the primary bottleneck is database throttling.

Tip: use batching + RU rebalancing before adding new hardware.

Impact timeline

From platform engineering to research labs, I like projects with real stakes.

Aug 2025 - Present
Teaching Assistant, Cloud Computing

Supported labs on virtualization, distributed systems, and AWS architecture.

Distributed systems Cloud labs
May 2025 - Aug 2025
Software Developer Intern, Paycom

Shipped a GraphQL CRM layer that reduced payload size and improved developer velocity.

GraphQL API design
Jul 2022 - Jul 2024
Software Engineer, Walmart Global Tech

Built authentication layers, performance improvements, and cloud migrations for Mexico e-commerce.

Performance Kubernetes Cloud migration
May 2019 - Jan 2020
Research + analytics projects

Human-bot behavior analysis on 0.6M X interactions and productivity analytics with clustering.

ML research Behavior analysis

How this site is built

A transparent look at architecture, tradeoffs, and what I would improve at 10x traffic.

Architecture

Client CDN Edge cache Metrics LLM demos

Static-first delivery with progressive enhancement and observable metrics.

Tradeoffs

  • Keep JS tiny; favor CSS for motion.
  • Prototype LLM demos locally to avoid latency.
  • Use real metrics later via a lightweight API.

At 10x traffic

  • Add edge caching for prompt results.
  • Move metrics to a read-optimized store.
  • Introduce SLOs and synthetic checks.

Lets build something precise and bold.

If you care about systems that feel quietly powerful, we will get along.

alisharaj2607@tamu.edu LinkedIn