work

hi there 👋 my name is taha. if we haven’t met yet, i hope this gives you a quick proxy for my skills and experience.

i’m an engineer with over 6 years of professional experience leading teams and projects in data and ml field. i am interested in building scalable and high-performing cloud based machine learning solutions. i’ve also been a founder and deeply familiar with how startups operate in different stages. currently I’m pushing LLMs to their edge to discover possibilities.

sabbatical ⚓️ _{^{(2023-08 to now)}}

~ built Practical.sh
~ built prototypes for Glu – wrapping APIs using LLMs, at scale
~ interviewed at Y Combinator, building infrastructure tooling(docker/terraform/kubernetes) on LLMs.
~ organized the Toronto Code Golf Tournament, with participation of engineers from Microsoft, Amazon, and other startups.
~ working on ai safety projects

data & ml @ Flinks _{^{(2022-10 to 2023-08)}}

i lead data infra and mlops –

~ supported +150 people with their data needs
~ lead migration of ml pipelines from AWS to GCP
~ lead the development of our largest ml pipeline streaming billions of predictions (team effort of 9 engineers and data scientists ) ~ built the data infra which powers most of data at flinks today

[Flinks raised a $103M series B]

Tools:

general tools: Python, Pytorch, PySpark, Bash, Airflow
infrastructure: Terraform, Docker, Kubernetes, Azure DevOps (CI/CD)
data tools: dbt, great_expectations, lightdash
GCP: VertexAI(Kubeflow), BigQuery, Dataflow (Apache Beam), Cloud Functions, Cloud Composer (Airflow), PubSub, Cloud Storage, MemoryStore (Redis), FireStore
AWS: SageMaker, Lambda Functions, S3, Redis, API Gateway, CloudWatch 

software engineer @ AI Research Lab _{^{(2020-09 to 2021-03)}}

supported a team of researchers building a multimodal generative ml engine . wrote code for preparing large datasets, training and fine-tuning various models, as well as serving them:

• fine-tuned GPT2 models on GCP GPUs
• trained image captioning models with PyTorch
• trained GANs for image generation
• generated melodies based on text sentiment
• wrote a REST API for serving these models, and a React Native app as an interface

Tools:
Python, Bash, Docker, PyTorch, GCP, Flask, React

data @ Banimode.com _{^{(2017-11 to 2018-04)}}

helped marketing team efficiently ramp up traffic by ~ 3x in 6 months. lots of python, sql and building stats models. banimode is one of the biggest online clothing stores in Iran.

product & marketing @ Poloto _{^{(2016-08 to 2017-09)}}

we built an ecommerce platform in iran, similar to shopify i worked between product and marketing, and talked to a lot of customers! also helped raise a seed round

.:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:~:..:*~*:..

education

~ cs and math @ york university

~ business @ sharif university

~ in high school i won a silver medal in a national physics olympiad.

~ i then ranked 405 among ~90k participants in a national math and physics competition in iran.

i gained most of my hackings skills from hacking hardware for many years. in university i skipped classes and spent most of my time writing low level sofware for ARM microcontrollers, freelancing software projects to make money, and taught linux to undergrads.

.:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:*~*:..:~:..:~:..:*~*:..