Started in 2021, Coram.AI is building the bestbusiness AI video system on the market. Powered by thenext-generation video artificial intelligence, we deliverunprecedented insights and 10x better user experience than theincumbents of the vast but stagnant video security industry. Ourcustomers range from warehouses, schools, hospitals, hotels, andmany more, and we are growing rapidly. We are looking for someoneto join our team to help us scale our systems to meet the userdemand and to ship new features. Team you will work with Founded byAshesh (CEO) and Peter (CTO), we are serial entrepreneurs andexperts in AI and robotics. Our engineering team is composed ofindustry experts with decades of research and experience from Lyft,Google, Zoox, Toyota, Facebook, Microsoft, Stanford, Oxford, andCornell. Our go-to-market team consists of experienced leaders fromVerkada. We are venture-backed by 8VC + Mosaic, revenue-generating,and have multiple years of runway. Being part of our team meanssolving interesting problems at the intersection of userexperience, machine learning and infrastructure. It also meanscommitting to excellence, learning, and delivering great productsto our customers in a high-velocity startup. The role We are hiringa Machine Learning engineer. - Take an existing open-source Pytorchmodel, fine-tune, productionize them in C++ runtime, and optimizefor latency and throughput. - Take an open-source model andfine-tune them on our in-house data set as needed. - Designthoughtful experiments in evaluating the tradeoffs between latencyand accuracy on the end customer use case. - Integrate the modelwith the downstream use case and fully own the end metrics. -Maintain and improve all existing ML applications in the product. -Read research papers and develop ideas on how they could be appliedto video security use cases, and convert those ideas to workingcode. Requirements - You should be a good software engineer whoenjoys writing production-grade software. - Strong machine learningfundamentals (linear algebra, probability and statistics,supervised and self-supervised learning). - Keeping up with thelatest in deep learning research, reading research papers, andfamiliarity with the latest developments in foundation models andLLMs. - (Good to have) Comfortable with productionizing a Pytorchmodel developed in C++, profiling the model for latency, findingbottlenecks, and optimizing them. - Good understanding of dockerand containerization. - (Good to have) Experience with Pytorch andPython3, and comfortable with C++. - (Good to have) Understandingof Torch script, ONNX runtime, TensorRT. - (Good to have)Understanding of half-precision inference and int8 quantization.What we offer - Company equity % in an early-stage startup.#J-18808-Ljbffr
