Nanonets is automating document information extraction using AI. We are headquartered in San Francisco. We are backed by prestigious investors from bay area like Y-Combinator, SV Angels, Sound Ventures by Ashton Kutcher. We are currently profitable and growing at a fast pace and looking to expand our team.

We are building a product that lets companies automate extracting key information from documents like invoices, receipts, or any other kind of document and integrate it into their workflows saving manual work. We need to keep building features that will let users automate millions of documents of different kinds every day, feed them to our AI for learning, plug our API to external systems like salesforce, quickbooks, RPA providers etc.

You should check it out at https://app.nanonets.com

Nanonets is a startup headquartered in the San Francisco Bay Area, solving real-world business problems with cutting-edge deep learning. We are backed by prestigious investors from Silicon Valley, such as Y-Combinator (Sam Altman was our group partner at YC), SV Angels, and Elevation Capital. Our product automates complex business processes involving unstructured data, using deep learning to convert it into a structured format and connect multiple applications with each other, all in an automated manner.

Since 2021, we have been building and using large-scale multimodal architectures in deep learning, such as GPT-4, which have gained popularity in recent times. Some of the recent work we are doing involves using these architectures to automate building workflows that will completely replace RPA as an industry.

If you are looking to work at a startup with really smart colleagues, working on state-of-the-art deep learning architectures, solving real-world problems, and have product-market fit with rapidly growing customers/revenue, Nanonets would be an ideal place for you!

Job Description

The role can be summed up as building and deploying cutting edge generalised deep learning architectures that can solve complex business problems like converting unstructured data into structured format without hand-tuning features/models. You are expected to build state of the art models that are best in the world for solving these problems, continuously experimenting and incorporating new advancements in the field into these architectures.

What We Expect From You

• Strong Machine Learning concepts
• Strong command in low-level operations involved in building architectures like Transformers, Efficientnet, ViT, Faster-rcnn, etc., and experience in implementing those in pytorch/jax/tensorflow
• Experience with the latest semi-supervised, unsupervised and few shot architectures in Deep Learning methods in NLP/CV domain
• Strong command in probability and statistics
• Strong programming skills
• Have previously shipped something of significance, either implemented some paper or made significant changes in an existing architecture etc

Interesting Projects Other Senior DL Engineers Have Completed

• Deployed large scale multi-modal architectures that can understand both text and images really well
• Built an auto-ML platform that can automatically select best architecture, fine-tuning method based on type and amount of data
• Best in the world models to process documents like invoices, receipts, passports, driving licenses, etc
• Hierarchical information extraction from documents. Robust modeling for the tree-like structure of sections inside sections in documents
• Extracting complex tables — wrapped around tables, multiple fields in a single column, cells spanning multiple columns, tables in warped images, etc.
• Enabling few-shots learning by SOTA finetuning techniques

Some of the interesting things our backend team has shipped

• Compile python code into C which could be imported into golang and then shipped as binary for on premise systems
• Autoscale GPU dependent services with kubernetes with a custom metric
• Displaying machine learning metrics in simplified ways to end users so they can act based on those metrics
• Building large number and variety of integrations with relatively generic interface like salesforce, quickbooks, RPA's, external databases
• Process large number of files in highly distributed manner in golang

Some of the interesting things our frontend team has shipped

• Ability for users to annotate documents so AI can learn which fields to extract
• Displaying machine learning metrics in simplified ways to end users so they can act based on those metrics
• Letting users build complex visual workflows around our API in our product.
• Let users visualize complex ML metrics in a very simple and intuitive way

Our stack:

• Databases
  • Cassandra DB
  • Postgres/MySQL
• Backend
  • Golang for API and other microservices
  • Python for Machine learning (Tensorflow, Pytorch)
• Frontend
  • React, Typescript
  • Mobx
• Cloud Providers
  • AWS
  • GCP for ML heavy workload
• Monitoring/Alerting
  • ELK for logging
  • Prometheus for Monitoring
  • Graphana for dashboards
• Orchestration
  • Kubernetes
• DevOps
  • Jenkins for CI/CD