Deep Learning Engineer

Location: Bangalore (Hybrid) | $40M+ Funded | Building State-of-the-Art AI

Nanonets is transforming the way businesses work. Our AI platform takes the manual, messy, time consuming work — that bog down industries like finance, healthcare, supply chain, and more — and turns them into seamless, automated processes. What once took hours of human effort now takes seconds with Nanonets. Our client footprint spans across 34% of Fortune 500 enabling businesses across various industries to unlock the potential of AI in automating their business processes.

More than 10,000 businesses trust Nanonets because we don’t just promise efficiency — we deliver it with unmatched accuracy, seamless integrations.

In 2024, we raised a $29M Series B led by Accel with continued backing from Elevation Capital and YCombinator, fueling our mission to reshape entire industries through intelligent automation. With revenues tripling year over year and a rapidly scaling global team, we’re not just imagining the future of work — we’re building it.

Read about the release here:

Article 1

Article 2

We’re on a mission to hire the very best and are committed to creating exceptional employee experiences where everyone is respected and has access to equal opportunity.

About the role

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’re looking for

• 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.
• 1-3 years of 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.

Ideal candidate should have the following skillset

• Python
• Tensorflow
• Experience building and deploying systems
• Experience with Theano/Torch/Caffe/Keras all useful
• Experience Data warehousing/storage/management would be a plus
• Experience writing production software would be a plus
• The ideal candidate should have developed their own DL architectures apart from using open source architectures.
• Ideal candidate would have extensive experience with computer vision applications.

Interesting Projects Other DL Engineers Have Completed

• Setting New Standards:

  Through our Automation Benchmark, we are defining how AI systems are measured on grounding, reliability, and performance.

• Proven Adoption:

  Our Nanonets-OCR-S model on Hugging Face has already

  ~225,000 downloads

  , validating its global impact and utility.

• Global Recognition:

  Our research and open-source contributions are recognized by leading voices in AI (example).

• Enterprise-Ready AI:

  Our models don’t just output predictions - they provide

  grounded answers with confidence scores

  to enable trustworthy decision-making.

• Agentic OCR Systems:

  Unlike traditional OCR, our models are

  agentic

  - capable of reasoning about inputs, adapting to task context, and chaining multiple steps to deliver structured, actionable data.

• VLM + LLM Innovation:

  From text to vision-language, we are solving alignment, hallucination reduction, and cross-modal understanding at scale -

  leveraging the latest techniques like RLHF, PEFT, and advanced fine-tuning

  to push what’s possible.

Key Responsibilities

• Understand specific customer requirements, develop and apply SOTA GenAI solutions to their workflows
• Develop and fine-tune OCR and Vision Language Models (text detection, recognition, entity extraction, layout understanding).
• Build and maintain data pipelines for documents, including cleaning, augmentation, and annotation.
• Implement and evaluate document parsing solutions for invoices, receipts, IDs, contracts, forms, etc.
• Work with LLMs/VLMs to enhance document understanding and enable intelligent reasoning over documents.
• Collaborate with senior engineers to deploy models in production with scalable APIs and workflows.
• Track and improve accuracy, robustness, and latency using proper evaluation metrics

Qualifications

Must-Have:

• 1–3 years of experience in Machine Learning / AI Engineering/ Deep Learning
• Strong programming skills in Python and familiarity with PyTorch or TensorFlow.
• Experience with data preprocessing, training, and evaluation for vision or NLP tasks.
• Experience working with LLMs or multimodal models (Hugging Face transformers, Nanonets-OCR-S, Qwen-VL, LLaMa).
• Knowledge of REST APIs, Docker, Git, Kubernetes and basic cloud deployment (AWS/GCP/Azure).
• Good understanding of ML fundamentals (supervised learning, evaluation metrics, error analysis)
• Basic understanding of agentic AI workflows (document reasoning, confidence scoring, grounding).
• Have previously shipped something of significance, either implemented some paper or made significant changes in an existing architecture etc.
• Strong problem-solving and analytical skills