Remidio is on a mission to bring cutting-edge healthcare access to everyone. We do this by building next-generation medical devices — combining custom hardware and AI into tools that anyone, anywhere can use.
Today, our products deliver some of the world's highest accuracies in eye disease detection and work with minimal infrastructure — without internet, medical specialists, or even electricity. This means any frontline worker can deliver expert-grade diagnosis anywhere — from the tops of remote mountains, to a rural village without proper health infrastructure.
Our technology has screened over 16 million patients across 55+ countries. It's used by India's largest healthcare institutions, government screening programs, and at major universities including the University of California and the University of Virginia.
Our work has been independently benchmarked by the National Health System of the UK, publicly recognised by global healthcare leaders such as Kiran Mazumdar-Shaw, and featured on the Emmy-winning medical drama The Pitt. More recently, our technology was endorsed and featured by Bill Gates, who highlighted and recognised its ability to save lives.
Our next frontier goes further: using the eye to find disease anywhere in the body. The retina is the only place where blood vessels can be observed directly and non-invasively. Our roadmap includes detection of heart disease, kidney disease, and pregnancy complications from a single eye scan. We believe this has the potential to become the most powerful diagnostic tool in medicine.
This work needs strong, curious people who care about the work actually reaching the people it's built for. If this resonates with you, we welcome your application.
We are hiring an AI Engineer to build state-of-the-art models that detect disease from retinal images — both eye conditions and diseases that extend across the body.
Your work will involve training models that go from a Jupyter notebook to a smartphone in a frontline health worker's hand. This is hands-on: design experiments, iterate on architectures, work directly with proprietary clinical datasets, and ship models that actually reach patients.
We hire for depth of understanding, not for which libraries you've used.
-
Math and intuition: You can explain backpropagation and gradients from scratch. You can reason about why a transformer outperforms an LSTM on sequence tasks, what an optimizer like SGD, Adam, or AdamW is doing under the hood, and why a particular architecture suits a particular problem.
-
Training models well: You've trained models end-to-end. You can read a loss curve, diagnose why training is unstable, and pick the right regularization, learning rate schedule, and batch size for the situation.
-
Dataset intuition: You understand that model quality starts with data quality. You can look at a dataset, spot class imbalance and label noise, and design splits and augmentations that don't leak.
-
Smart, fast learner: Whatever specific skill we list below that you don't have today, you're confident you can pick up.
-
Python fluency: Clean, readable code.
-
Design, train, and ship deep learning models for retinal image analysis — across eye disease and broader disease detection from the same scans.
-
Work with proprietary clinical datasets, including handling messy real-world data, class imbalance, and multi-site variation.
-
Optimize models for on-device, offline inference on smartphones (quantization, pruning, distillation).
-
Collaborate with clinical, regulatory, and product teams to translate model performance into validated medical impact.
-
Stay current with vision and foundation model research, and bring relevant ideas into our work.
-
Bachelor's or Master's in CS, EE, Math, Physics, or a related field.
-
Hands-on experience in building models. We care about what you've actually built and shipped — projects, publications, open-source contributions, or strong references — much more than years on a CV.
-
Computer vision experience (CNNs, ViTs, segmentation, multi-task heads).
-
On-device or efficient inference (quantization, pruning, mobile deployment).
-
Foundation models or self-supervised pretraining.
-
MLOps and deployment experience — we'd like you to grow into this, but it's not required to start.
-
Healthcare or medical imaging exposure.
Your work will draw on multi-site, real-world clinical data — from our own deployments and from partnerships with credible healthcare institutions. You'll be working with real data, not curated benchmarks.
You'll have end-to-end ownership across the full pipeline — architecture, training, clinical validation, and deployment.
You'll work on technically demanding and consequential problems at scale. Engineers who take full ownership develop deep expertise through the challenges they solve here.
Our team has published in JAMA and Nature, presented at top healthcare conferences, and partnered with institutions like the Gates Foundation. You'll have the opportunity to do the same — and to hold patents in your name.
