mm.
Software engineerNew York, NY

Murtaza
Mister.

Murtaza Mister outdoors in winter, smiling

Curious about the why.
Careful with the how.

I build the backends, pipelines, and distributed systems that production runs on, and I stay with them until they hold.

Software Development Engineer at Scorealytics, leading engineering on a production risk intelligence platform.

See the impact See the work
01 / Experience

From research
to production.

Owning platforms, raising the bar,
and shipping under real load.

Jun 2026 to present

Scorealytics

Software Development Engineer

Lead engineer on client-facing projects for a regulatory risk intelligence platform. I own the risk-scoring service end to end: architecture, the ingestion pipeline, production releases, and the incidents that follow.

Designed a crash-safe work queue that runs at 99.8% job success with zero lost or duplicated work. Scored 5,000 regulations in 48 minutes with zero failures. Grew the test suite 9x to 238 automated checks, shipped 175+ commits across 7 services in four months, and took a client-facing risk report from commitment to delivery in 8 days.

Python · TypeScript · Next.js · Node.js · SQL · MongoDB · Azure · LLMs

165K+

risk scores recomputed in 48 hours,
peaking at 4,500+ an hour, during
a live model upgrade

250K+

live risk scores served by
the platform I own

99.8%

job success in production,
zero lost or duplicated work

Rounded production figures, June to October 2026.

Jan to Aug 2025

Research Foundation for SUNY

Graduate Research Assistant

Trained neural controllers to stand in for model-predictive control in U.S. Army-funded UAV-swarm research. Generated millions of training samples and coordinated 40+ simulated UAVs collision-free across obstacle avoidance, adversarial pursuit, and long-horizon navigation.

Ran parallel training and simulation at scale on SeaWulf HPC with SLURM, Python, TensorFlow, and MATLAB.

Dec 2023 to Mar 2024

ISRO

Software Engineering Intern

Built a real-time rover telemetry stack across React, Node.js, WebSockets, and PostgreSQL. Fused GPS, NavIC, and pseudolite measurements with a Kalman filter, cutting positioning error by 35%.

May to Jul 2023

Mastercard

Software Engineering Intern

Built Java and Spring Boot services that migrated 100+ GB of financial data with zero downtime and moved 750+ schemas, using parallel partitioning and batched I/O. Added a React/Redux dashboard so operators could watch progress, retries, and errors live.

Shipped with Oracle SQL, Jenkins, Docker, and Kubernetes.

May to Jul 2022

Simulas

Software Engineering Intern

Built 25+ APIs and webhooks for property-management workflows. Reworked MongoDB aggregations and indexes, taking query latency from 1.7 s to 250 ms, over 85% faster.

Also built Angular analytics and AWS infrastructure with Terraform.

02 / Work

Built to find
out how.

Consensus, concurrency, real-time data,
and applied AI. All running code.

01
Distributed systems / Graduate coursework

Fault-tolerant banking

What happens to a transaction when a server fails halfway through, or starts lying? Replicated banking systems that keep money consistent through crashes, malicious nodes, and cross-shard transfers.

  • 12replicas, Byzantine-fault tolerant
  • 15Ktransactions benchmarked
  • ~200TPS through leader changes
Design notes

Separate implementations use Paxos for crash faults and PBFT with signed messages for Byzantine faults. Two-phase commit keeps cross-shard transfers atomic, and write-ahead logs drive recovery. Benchmarks held around 120 ms latency. The public repositories cover the core consensus implementations.

Java · Spring Boot · Paxos · PBFT · Two-phase commit

Paxos code ↗PBFT code ↗
02
Applied AI / Mobile

PantryPal

An AI kitchen assistant that keeps your pantry accurate on its own. Tell it what you cooked in plain language, and an LLM turns that into a structured, editable list of ingredient deductions. It also generates recipes from what you actually have, so less food goes to waste.

  • 100%of model output schema-validated
  • 6kitchen steps closed in one loop
  • Secondsfrom cooking log to updated pantry
Build notes

OpenAI runs only behind backend endpoints with strict JSON contracts, so keys stay server-side and every response is normalized before it touches inventory. Recipe chat refines a dish, then finalizes it into a deduction draft. Overuse validation and undo guard every pantry change. Shopping list, pantry, recipes, cooking logs, and reminders all live in one flow.

OpenAI · React Native · Expo · TypeScript · Express · Supabase · Redis

Demo video ↗Source code ↗
03
Concurrency / Backend

PhotoPort

A concurrent image crawler with bounded worker pools, duplicate prevention, same-site discovery, and exponential backoff. Streams discovered images to the browser as they are found.

  • 0duplicate page fetches
  • Adaptiverate limiting with backoff
  • Liveresults streamed while crawling
Design notes

A fixed worker pool runs concurrent crawls while the response streams results. Crawl-depth and concurrency limits bound the work, and domain checks keep discovery on the requested site. The repository is named ImageFinder.

Java · Jetty · Jsoup · Maven · JavaScript

Source code ↗
04
Data visualization / Web performance

Jetstream Tracker

Compare weather-balloon trajectories with forecast winds on an interactive globe. Background processing and spatial indexing keep a huge, moving dataset smooth.

  • 10K+live trajectories on one globe
  • 60FPS while panning and playing back
  • <200msinteraction latency
Design notes

Web Workers process trajectories away from the interface. RBush spatial indexing and viewport culling limit rendering work to the current map view, with 24-hour playback, lazy loading, and caching.

React · TypeScript · MapLibre GL · Web Workers · RBush

Live demo ↗Source code ↗
05
Real-time systems / Full-stack

Resonate

Shared music playback across devices, with rooms and real-time coordination. Keeps everyone on the same beat even when their clocks and networks disagree.

  • <150msplayback drift across devices
  • 300+concurrent clients in load tests

JavaScript · React · Node.js · MongoDB · WebSockets

Source code ↗
More projects Research, experiments, and earlier work
06

pdfRAG

Ask your PDFs a question. Retrieval-augmented answers stream back with page-level citations, backed by Qdrant vector search.

Code ↗
07

Neural Flocking

Neural controllers for UAV swarms, trained on millions of samples with HPC. Research at SUNY.

Research ↗
08

BreakNeck

Earable head-gesture recognition for hands-free mobile control.

Code ↗
09

PongSense

Hands-free Pong with computer vision and speech input.

Code ↗
10

CareConsole.ai

Hackathon project for symptom tracking and patient dashboards.

Code ↗
11

vocalForLocal

A marketplace for local businesses.

Code ↗
12

Connect

Real-time chat with shared and private rooms.

Code ↗
13

Conspectus

Webpage summarization and highlighting.

Code ↗
14

LifeCart

An earlier Java backend project.

Code ↗
15

Sales & inventory management

C# and .NET MVC application for managing inventory.

Code ↗
16

Employee Management System

An ASP.NET Core MVC application with Entity Framework.

Code ↗
17

GRE Vocab Game

A vocabulary-learning project built with Dart.

Code ↗
18

Competitive Trainer

A tool for selecting programming-practice problems.

Code ↗
19

BookListRazor

A Razor Pages and Entity Framework CRUD application.

Code ↗
20

ToDo

An early JavaScript task-list app.

Code ↗
21

osFence

A kernel runtime policy-engine project. No public repository.

22

Lucidify

Image dehazing with convolutional neural networks. No public repository.

23

Planscape.ai

An agent-assisted planning project. No public repository.

03 / About

Follow the
problem all
the way down.

My favorite moment in engineering is when a confusing failure finally makes sense.

A slow query, a queue that stops draining, a consensus round that never commits. I find the cause, make a clear call, and get the fix into people's hands. Then I make sure it cannot happen the same way twice.

I taught Distributed Systems as a teaching assistant at Stony Brook and, as president of the Computer Society of India branch at DDU, ran initiatives for 400+ students.

Education & coursework

Stony Brook University

MS, Computer Science 2024 to 2026

GPA 3.9 / 4.0

  • Distributed Systems
  • Analysis of Algorithms
  • System Security
  • Human-Computer Interaction
  • Simulation and Modeling

Dharmsinh Desai University

BTech, Computer Science 2020 to 2024

GPA 3.82 / 4.0

  • System Design
  • Software Engineering
  • Web Services Development
  • Database Management Systems
  • Operating Systems
04 / Toolkit

What I reach for.

Backend & data

  • Java
  • Spring Boot
  • Express
  • PostgreSQL
  • Redis
  • FastAPI
  • WebSockets
  • C++
  • C#
  • Fastify
  • Oracle SQL
  • Spring Batch
  • REST APIs
  • ASP.NET Core
  • Entity Framework
  • Razor Pages
  • Maven
  • Jetty
  • Jsoup

Infrastructure & delivery

  • Docker
  • Kubernetes
  • Terraform
  • AWS
  • Azure DevOps
  • Jenkins
  • Linux
  • pytest
  • Playwright
  • Azure Key Vault
  • Azure Functions
  • Azure Table Storage
  • Application Insights
  • Entra ID
  • Jest
  • CI/CD

Web & mobile

  • HTML
  • CSS
  • Redux
  • Angular
  • MapLibre GL
  • Web Workers
  • RBush
  • React Native
  • Expo
  • Supabase
  • Zustand
  • zod
  • Dart

Applied AI & research

  • OpenAI
  • Azure OpenAI
  • Qdrant
  • RAG
  • TensorFlow
  • Keras
  • MATLAB
  • SLURM
  • NumPy
  • OpenCV
  • MediaPipe
  • Pygame
  • Vosk
  • SpeechRecognition
  • PyAudio
  • YAML
  • Jupyter
05 / ContactNew York, NY

Have something
in mind?

For the keyboard people

Shortcuts

g e
Go to experience
g w
Go to work
g a
Go to about
g c
Go to contact
b
Break the cluster (playground)
d
Toggle dark mode
t
Back to top
?
Show this panel