Kamil Arif.
Need an End-to-End CV pipeline?
Data Scientist at AT&T building optimization and anomaly-detection systems for a national network. I work where applied ML meets hard operational constraints — algorithms that have to survive contact with real infrastructure, real budgets, and real engineering agreements.
Commits over time
86 repos · since Jan 27, 2019
Experience
AT&T — Sr. Specialist, Member of Technical Staff
— Present
Artemis Network Planning
- Build algorithms that search for cost-effective L2/L3 network topologies subject to hard engineering and latency constraints.
- Finalized plans now cover regions servicing 12.4M people, with $60M in confirmed cost savings to date.
- Work directly with network operations at several levels to pin down real requirements and anticipate failure modes before they reach the plan.
AT&T — Data Analyst
—
IPSEA — Technical Development Program
- Developed an IP reallocation algorithm with $9M in theoretical savings, freeing over 300K addresses.
- Refactored the team's data pipelines for better parallelization and use of provisioned compute.
- Built anomaly detection over backbone DNS server metrics to catch cyberattacks and infrastructure failures, and laid the foundation for the frontend team's dashboard.
NJIT — Grader
—
CS 370 — Intro to AI / Data Mining
- Graded assignments for a senior-level data mining course — 228 students across 3 semesters.
- Worked with a team of graders and the professor to keep standards consistent.
- Took an active role in helping students learn from mistakes and build real intuition for AI/ML concepts.
CDx Diagnostics — Intern
—
Cancer Cell Classification
- Differentiated cancerous cells from debris in dyed microscope slides, owning collection, annotation, training, and testing.
- Reached 89% detection accuracy on a limited dataset by fine-tuning a ResNet-50.
- Set up documentation for every deliverable at the end of the internship.
Selected work
All projects →ML Architecture & Topology Research
Independent study into novel ML architectures, measured by computational efficiency and knowledge representation.
- PyTorch
- Topology
- Visualization
Buckshot Roulette — RL in Stochastic Games
A Python engine for a hidden-information game, built to train reinforcement learning agents against it.
- Python
- Reinforcement Learning
- Tensorboard
GalaxyGen
A worldbuilding and mapping tool for sci-fi galaxies — procedural generation with an interactive editor.
- TypeScript
- Next.js
- Pixi.js