saberhq.com
Friendly notes on what I’m building, reading, and figuring out.
Explore why autoencoders shine; this post pairs plain-language basics with demos, figures, and a walkthrough video.
Read more →
ntEmbd turns nucleotide sequences into vector embeddings; explore training setup, benchmarks, and GitHub repo links.
Read more →
Trans-NanoSim captures nanopore transcript quirks; learn how benchmarks and simulations accelerate RNA tooling workflows.
Read more →
My name is Saber and I’m a research scientist with years of experience in genomics, data science, and machine learning. For the past year at Genentech (gRED), I co-led the analysis of two genome-wide, multi-million-cell single-cell CRISPR Perturb-seq screens. Because every design choice in a Perturb-seq pipeline — QC thresholds, confounder correction, statistical modeling, dimensionality reduction — changes the biology you end up inferring, I also built tooling to make those consequences visible: a CLI that renders interactive dashboards comparing outcomes across parameter sweeps, plus a sweep orchestrator for Nextflow pipelines on HPC.
More recently, I have been building Sidechain, my solo entry in Arc’s Virtual Cell Challenge 2026. My first paper (Nucleic Acids Research, 2016) modeled how RNA-binding proteins and microRNAs jointly govern transcript fate, and Sidechain is my bet that this post-transcriptional layer — written in sequence, and therefore stable across cell contexts — is a prior most perturbation-response models leave on the table.
I earned my Ph.D. in Bioinformatics at the University of British Columbia (UBC), where I was advised by Prof. Dr. Inanc Birol, working at the Bioinformatics Technology Lab. During my Ph.D., I broadly worked on developing computational tools and software solutions for next-generation long-read sequencing technologies. My doctoral dissertation (see here) was focused on utilizing machine learning in transcriptome analysis, and my time at the BC Cancer Genome Sciences Centre produced ntEmbd, a deep learning embedding model for nucleotide sequences, and NanoSim, a long-read simulation suite the field uses for benchmarking (62,000+ downloads). Before starting at UBC, I received my M.Sc. in Bioinformatics from METU, where I was advised by Dr. Hilal Kazan and Dr. Yesim Aydin Son. My B.Sc. was in Information Technology Engineering.
This personal website is my notebook in public, where I share my journey in personal life and professional career. Say hi and stay in touch :)