Sachin Palahalli Chandrakumar Email & Phone Number

Burlington, Massachusetts, Us

1. Developed and deployed a chatbot using the Mistral Instruct 7B model, with Ray for API management, enabling efficient handling of asynchronous requests.2. Achieved real-time performance by deploying the chatbot with Multi Core vLLM, leveraging parallel processing over dual NVIDIA 4090 GPUs to handle 80 requests per minute during stress testing.3. Optimized neural network model training by parallelizing training computations across multiple GPU’s.4. Achieved a 60% reduction in model training duration by effectively maximizing GPU parallel computations.5. Integrated Active Learning with Reinforcement Learning to achieve best accuracy with minimum number of labeled data.

Boston, Ma, Us

Assisted 80 students in understanding Information retrieval, machine learning concepts and helped them with coding and debugging in python and java.Created assignments and coding exercises based on Information Retrieval , machine learning concepts for the students to better understand the subject.

Frankfurt Am Main, Hessen, De

1. Integrated BERT embeddings to analyze unstructured text data in SWIFT messages for Named Entity Recognition(NER) and anomaly detection.2. Built an ensemble model combining unsupervised clustering(DBSCAN, Isolation Forest) with supervised models like XGBoost and deep-learning LSTM model and improved accuracy by 30%.3. Spearheaded the development of microservices to handle post-transactional operations of the payments processed.4. Efficiently optimized data retrieval time by 60%(5 minutes to 2 minutes) by writing efficient SQL queries and database indexing.

Mumbai, Maharashtra, In

1. Utilized machine learning and pattern recognition to identify complex fraud patterns, enhancing fraud detection system accuracy.2. Streamlined fraud verification process with ML algorithms, reducing verification time by 90% (5 minutes to 30 seconds).3. Led a 3-member team in the design and development of a web application for support team to access sensitive data securely.4. Boosted performance by implementing an event listener to publish events into Google Cloud Pub/Sub, resulting in a 30% increase in data processing speed.Tech Stack: Python, Linux, Scikit-learn, ML, Deep Learning, TensorFlow, Elixir, MySQL, PostgreSQL, GitHub, Java, Docker

Bangalore, Karnataka, In

Developed a backend web application with Restful web services in java using Spring Framework. Led and developed UI automation using Protractor which is an end to end behavior-driven testing framework.

Master Of Science, Artificial Intelligence

Master Of Engineering - M.E, Computer Science

Bachelor Of Engineering - B.E, Computer Science And Engineering