01Key Responsibilities
Design, develop, and deploy machine learning and deep learning models for real-world applicationsPerform statistical analysis and data exploration to derive meaningful insights (Time series analysis & Forecasting, predictive analysis)Build and optimize NLP and Computer Vision pipelines for various use casesFine-tune transformer-based models for tasks such as text classification, entity recognition, and vision tasks (BERT, LayoutLMv3, Vision Transformers,etc)Develop and enhance OCR systems, including model fine-tuning and post-processing pipelines (LayoutLMv3, PaddleOCR, DocTR, Surya)Build document AI systems (text + layout + image)Perform feature engineering and EDAWork on domain-specific keyword extraction systemsContribute to R&D in NLP, OCR, and multimodal AIOptimize models for accuracy, scalability, and performanceCollaborate with cross-functional teams to integrate ML models into production systemsContinuously evaluate and adopt new AI/ML techniques, tools, and frameworksMentor and guide junior engineers, helping them understand core concepts and best practicesMaintain documentation and ensure knowledge sharing within the teamRequired Skills & Experience:45 years of hands-on experience in AI/ML developmentStrong foundation in:ML, Deep Learning Statistics & Mathematics Expertise in:NLP + Computer Vision Hands-on with:Transformer models (e.g., BERT, Vision Transformers) and fine-tuning techniquesOCR systemsExperience with OCR technologies and model optimizationProficiency in Python and ML/DL frameworks such as TensorFlow /PyTorch Experience in building end-to-end ML pipelines and Strong algorithmic thinkingUnderstanding of data preprocessing, feature engineering, and model evaluation techniques Strong problem-solving skills and ability to work on R&D-focused tasksNice to Have:Experience with Generative AI and deep learning architectures (LLMs, diffusion models, etc.)Model optimization (quantization, pruning)Familiarity with MLOps practices and model deployment workflowsExperience with cloud platforms (Azure, AWS, or GCP) Knowledge of vector databases and embedding techniques Experience with annotation tools and dataset creation for OCR/NLP tasksExposure to model optimization techniques (quantization, pruning, distillation) .
