Machine Learning Engineer (North America Remote OK)

Company, Clients & TeamCompanies & Organizations own over a Billion IP cameras around the world. curbFlow's mission is to use computer vision to convert them to automated business intelligence machines.
Clients connect their video to curbFlow to get real-time critical data like pedestrian counts for retail, guest tracing for commercial real estate, traffic counts for DOTs and engineering firms, employee engagement at restaurants, and live wait times for amusement park rides, to name a few use cases.
Paying clients include Taco Bell/Yum Brands, NYC's largest 7-11 franchisee, Goodwill, Brightline, Merlin Entertainments (owner of Legoland) and JBG Smith, developer of Amazon's HQ2.
curbFlow was seed-funded with $8mm in 2018 by venture capital firms General Catalyst and Initialized Capital (Garry Tan) and expects to raise a Series A by the end of 2022.
The company was founded by Ali Vahabzadeh, who founded Chariot and sold it to Ford Motor Company for $65mm in 2016.
About the RolecurbFlow is searching for talented ML engineers with a track record of high achievement who can take responsibility for the full software development lifecycle, including:1) Building scalable pipelines for our model retraining stack2) Integrating models into our existing perception stack3) Integrating new tools into our embedded and cloud platforms

Role & Responsibilities

• Build scalable Machine Learning Pipelines that can be used to train and deploy object detection models for each of our clients on a consistent basis
• Build and operate our image data pipeline for collection, labelling and re-training our object detection models
• Build algorithms to understand what images need to be labelled from our data warehouse
• Build software libraries that standardize the acquisition, ingestion and integration of images for training object detection models
• Build pipelines that trigger retraining, pruning and deployment of object detection models after each labelling job is completed
• Establish and maintain the company's data lake/data warehousing strategy, define the appropriate data architecture, implement the best technical solution, and continue to meet the growing needs of the business
• Work with product and business analytics teams to ensure availability and accessibility of relevant business data and business metrics for product analytics and business performance reporting
• Build software libraries that standardize the acquisition, ingestion and integration of images for training object detection models for our customers
• Design and develop scalable platforms and processes for feature extraction, model training, and simulation
• Own tools, processes and controls to help the team grow at scale

Qualifications you Should Have

• 2+ years of professional Python experience/Expertise training, evaluating, and deploying models with a deep learning framework such as PyTorch or Tensorflow, Keras, Lightning, etc
• Familiar with computer vision techniques (i.e. object detection, object tracking, classification, image segmentation, etc.)
• Experience with OpenCV or similar libraries
• Experience with Linux environment and targeting embedded deployment
• AWS experience with S3, EC2, SQS, Lambda
• Excellent written and verbal communication skills
• Exceptional analytical skills
• Ability to operate and execute independently without significant oversight

Nice to Have's

• Deepstream and GStreamer experience
• TensorRT experience
• Experience working with embedded systems like Nvidia Jetson, Raspberry PI, etc.

Our first ML Engineer will be reporting directly to our CV Engineering Lead and be working on the front line of ensuring our client data is training on the most efficient models and ensuring peak accuracy.