AI/ML & Backend Engineer

New York · Open to full-time roles

I build intelligent systems that hold up in the real world.

I’m Mohammad Maaz Rashid, an engineer working across AI infrastructure, retrieval, and backend systems. I am currently researching high-update-rate vector search at NYU.

Explore my work ↓ View résumé ↗

Based inNYC

01 Systems thinker

02 Research → production

03 Software engineer

04 Tech enthusiast

99%task completion
for multi-agent routing

65%less manual
validation time

0.98F1 on financial
document routing

25K+customer records
automatically evaluated

AI INFRASTRUCTURE✦VECTOR SEARCH✦DISTRIBUTED SYSTEMS✦BACKEND ENGINEERING✦LLM EVALUATION✦ AI INFRASTRUCTURE✦VECTOR SEARCH✦DISTRIBUTED SYSTEMS✦BACKEND ENGINEERING✦LLM EVALUATION✦

Selected work

Projects where engineering depth meets practical impact.

Featured · 2026WebGL / Generative AI

ShaderLab

A generative GPU image editor in the browser.

Natural-language edits become validated GLSL shader passes. Every adjustment runs through a linear-space, premultiplied-alpha WebGL2 pipeline that feels instant while still rejecting broken shaders before they render.

Launch live project ↗

shader.frag

uniform sampler2D image;
uniform float intensity;

void main() {
  vec4 color = texture(
    image, v_uv
  );
  color.rgb = mix(
    color.rgb,
    pow(color.rgb, vec3(.8)),
    intensity
  );
}

60 FPS

02 / 03

Edge AI · 2025

Focus Assist

Thread-safe Pomodoro tracking with on-device NPU inference for Snapdragon X Elite laptops.

🏆 Winner, Qualcomm HaQathon

View repository ↗

03 / 03

Deep Learning · 2024

Lean ResNet

Reworked channels, filters, and skip connections to produce a sub-5M parameter classifier with 87% test accuracy.

View repository ↗

Experience

Building across the stack, from research prototypes to production platforms.

JUN 2026 TO PRESENT

Graduate Research Assistant

NYU Courant · Vector Databases & ANN Search

Developing VecFast, a two-tier ANN index combining an in-memory HNSW hot tier with a cluster-based cold tier and adaptive distribution-shift detection.

MAY TO AUG 2025

Machine Learning Intern

Qualcomm

Built RAG, evaluation, and six-agent orchestration systems for modem QA; reduced manual validation by 65%, LLM cost by 30%, and telemetry analysis time by 35%.

AUG 2022 TO JUN 2024

Software Engineer

Barclays

Shipped financial NLP, semantic search, and FastAPI services. Reached 0.98 F1 on document routing and automated evaluation across more than 25,000 customer records.

AUG 2025 TO MAY 2026

Graduate Teaching Assistant

NYU Courant · Computer Systems

Led weekly recitations for 60+ graduate students on CPU pipelines, memory hierarchy, virtual memory, and I/O systems.

How I work

I like owning the whole path, from data to deployment.

My work sits at the intersection of software engineering and applied machine learning. I’m most useful when a problem crosses boundaries: research and product, model and infrastructure, prototype and production.

At Barclays, that meant building FastAPI services alongside the Spark and AWS pipelines feeding them. At Qualcomm, it meant creating both the agent system and the evaluation harness that judged it. At NYU, I’m applying that same systems lens to vector search under sustained streaming updates.

2024 TO 2026 MS, Computer Science

New York University

2018 TO 2022 BTech, Computer Science

SRM Institute of Science and Technology

CURRENT TOOLKIT

PythonC++TypeScriptPyTorchTransformersHNSWFAISSPostgreSQLRedisFastAPISparkAWSDockerKubernetes

Have a hard systems problem?

Let’s build something that works.

I’m open to full-time software, ML, AI infrastructure, and data engineering roles.

mr7374@nyu.edu ↗