Los Angeles Area · Available for opportunities

I build machine learning systems

I’m Farhan, a software engineer working across LLM pipelines, backend infrastructure, and machine learning. I turn ambitious ideas into dependable, user-focused products.

  • LLM systems
  • Backend engineering
  • Machine learning
  • Full-Stack Development
Farhan Zaman smiling outdoors

About

Engineering with rigor

I’m a software engineer with a computer science background in machine learning and intelligent systems. I enjoy work that spans the stack: understanding the user, shaping the system, and getting the details right.

I’m most interested in applied AI, dependable infrastructure, and digital products that feel straightforward to use.

Working toolkit

Python · C/C++ · TypeScript · JavaScript · React · Next.js · REST APIs · PyTorch · AWS · Docker · MongoDB

Selected work

I build for real-world complexity

Production infrastructure, applied AI, and research systems designed around measurable outcomes.

Bar chart showing the class distribution of emotions in the children's drawings dataset
Computer vision research2025

Reading Emotion in Children’s Drawings

Compared ResNet-50 and Vision Transformer models for four-class emotion recognition, using semi-supervised learning to improve accuracy.

84%final accuracy, up from 78%
PythonPyTorchResNet-50Vision TransformerSemi-supervised learning
View repository
DeepLabV3 architecture used to identify electrified and unelectrified settlements from satellite imagery
ML for social impact2024

Mapping Energy-Scarce Settlements

Used satellite imagery and semantic segmentation to identify unelectrified settlements across Northern Africa and support NGO planning.

PyTorchDeepLabV3Satellite data
View repository

Experience

From models to products.

LifeSphere News logo

Full-Stack Developer

LifeSphere News

Jun 2025—Present · Los Angeles, CA

  • Built an LLM-driven news pipeline spanning article creation, validation, storage, and delivery.
  • Implemented quality controls for source verification, sensitive-topic handling, and duplicate detection.
  • Developed scalable backend APIs and internal evaluation tooling for production monitoring.
Mondego Research Group logo

Software Engineer Intern

Mondego Group, UC Irvine

Jun—Sep 2024 · Irvine, CA

  • Expanded and optimized UI components for a platform providing specialized AI assistants.
  • Architected login systems with database integration, REST endpoints, and OAuth authentication.
  • Built and deployed test and immigration law AI assistants and a retrieval-augmented generation framework.

Grader & Course Assistant

Donald Bren School of ICS

Jan—Jun 2024 · Irvine, CA

  • Designed Python-based grading infrastructure supporting more than 120 students.
  • Built scripts to evaluate student code against tests, edge conditions, and correctness criteria.
  • Automated spreadsheet aggregation and debugged submissions at scale to improve grading consistency.
Delvian Studio logo

Currently building

Delvian Studio

An independent software studio focused on custom applications and practical digital compliance solutions.

I’m developing the studio’s first service offerings and looking for thoughtful early partnerships where dependable software can make a meaningful difference.

Focus areas

  • 01Custom software
  • 02AI-enabled products
  • 03Digital compliance

Contact

Have an ambitious problem?

I’m open to software engineering opportunities and select early-stage client projects.

contact.farhanzam@gmail.com