About

Suvro Banerjee is a highly skilled Machine Learning Engineer and Engineering Manager with over 13 years of experience in software development and machine learning. He has a proven track record of building complex NLP pipelines using Deep Learning and taking them to production to solve business problems. Suvro has adhered to industry standard Coding, MLOps, and Optimization practices while model building, inference, and deployment at scale. He has built the Data Science/Machine Learning practices and team of researchers and engineers from ground-up for the organization. Suvro is highly passionate about Machine Learning and problem solving in general, both professionally and academically. He has an MSc in Econometrics with specialization in Optimization and Machine Learning from SUNY Buffalo 2013-15, where he was an excellent performer. Suvro's core skills include Deep Learning, Machine Learning, NLP, Transformer Based Models, MLOps, CI/CD, Systems Design, Google Cloud Platform (GCP), Computer Vision, Algorithms and Data Structures, Mathematics, and Optimization. He is proficient in Python, PyTorch, Hugging Face, ONNX, FastAPI, Docker, Cloud Build, Kubernetes, Cloud Logging, Message Queues (Pub/Sub), Asynchronous pipelines, Kubeflow, Google Cloud Platform, NVIDIA Triton Inference Server, TensorFlow, Apache Beam, Google Dataflow, Apache Spark, Big Data Stack (Hadoop), SQLs, PostgreSQL, NoSQL, Flask API, Uvicorn, Gunicorn, SAS, and R. Suvro is currently an Engineering Manager (Machine Learning) at Salesken, where he is building a conversational AI SaaS product to manage the end-to-end sales process for enterprises. As a founding member of the ML practice in Salesken, he has built and supervised various use cases, including a semantic search engine, a signal management system, an offline generation of cues, and an ML pipeline (MLOps). Before joining Salesken, Suvro worked as a Senior Data Scientist at Juniper Networks and a Data Scientist at Verizon, where he built various models to help the business solve operations problems, reduce costs, improve the digital channel mix, and devise market strategy with the help of Data Science. Suvro has also published a paper titled “Curriculum learning for adapting models on Code-Mixed Data in Emotion Recognition\" in COLING 2022 (International Conference on Computational Linguistics). He has completed various certifications and special