Abhishek Mukundan

Technology Services Group (TSG) @ MOODY’S | Delivery Services Team & Center of Excellence Team

LightLeap AI Team

- Improved web scraping efficiency by 25% by optimizing URL decoding techniques using Python and Beautiful Soup. - Refined vector database and file ID mapping, significantly boosting data retrieval efficiency by 35% using FAISS . - Led integration of our prototype into the LangFlow framework, reducing manual setup by 40% and streamlining workflows. - Fine-tuned GPT-4 models to improve response accuracy, enhancing user interaction quality by 25% for over 30 clients. - Managed AWS S3 and Lambda, achieving a 35% improvement in scalability and 40% cost reduction in operations

- Developed Python-based web scrapy, extracting 10,000+ data points from 10+ websites, boosting product listings by 30% - Improved PostgreSQL validation and Dagster pipelines, boosting data accuracy by 25% , cutting processing time by 40%. - Fine-tuned LLaMa2/7B LLMs, increasing user engagement by 20% through optimized marketing content generation - Collaborated with a cross-functional team to streamline data operations by 35%, boosting growth for 10+ local businesses

-Performed sentiment analysis, handling around 100GB of data and processing 200,000 rows to enhance sentiment classification using Hugging Face Sentiment Analysis models like XL-Net, RoBERTa, achieving a notable 7% boost over the 76% baseline for highly accurate predictions. -Employed both cased and uncased BERT models to fine-tune sentiment analysis specifically for article text, ensuring a more nuanced approach to data analysis within the Big Data spectrum. -Spearheaded end-to-end customer cycle initiatives for a Canada-based technology company in Marketing Analytics, driving a notable 25% increase in revenue per user and a 30% improvement in cross-selling opportunities within the retail landscape through hyper-personalized recommendations. Collaborated directly with clients, identifying key business needs, onboarding them onto our platform, and tailoring processes within their systems. This holistic approach enhanced customer retention and revenue streams significantly. -Shadowed a Marketing Analytics Unit project, leading advanced Customer Lifetime Value (CLTV) analysis on gaming sales data. This analysis resulted in a marked improvement in KPIs, strengthening customer loyalty to our product. Developed data-driven recommendations for strategic enhancements, ensuring a proactive approach towards sustained customer engagement and product optimization. -Devised RFM (Recency, Frequency, Monetary) scores from game sales data, optimizing customer segmentation. -Attained an exceptional 81% testing accuracy employed in Vertex AI, showcasing improved model performance.

-Developed advanced MLOPS methods utilizing MLFLOW within the Big Data landscape, contributing to a remarkable 40% improvement in algorithm precision. -Seamlessly integrated MLFLOW into the project, allowing for meticulous tracking of metrics including cohen kappa score and precision, optimizing MLOps for seamless end-to-end pipeline execution within the Big Data environment. -Orchestrated the management and storage of crucial metrics within MLFLOW, enhancing the evaluation and comparison process for informed decision-making. Leveraged models such as TensorFlow and Scikit-learn within the MLFLOW framework to achieve the desired cohen kappa score and precision metrics. This comprehensive approach ensured a deep understanding of model performance and guided decision-making processes effectively. -Employed Spark clusters to efficiently store and process real-time data streams, enabling the rapid ingestion and analysis of large datasets. This approach facilitated the seamless handling of data within the MLFLOW framework, ensuring smooth and efficient model development and deployment workflows within the Big Data environment.

-Orchestrated a pioneering internship experience, delving into the intricacies of Pairs Trading Algorithm and Basket Trading concepts, driving innovation in finance. -Applied advanced mathematical and statistical techniques to successfully implement these algorithms in real-time using MS Excel. -Performed thorough Time Series Analysis (TSA) on live data, refining analytical acumen and practical proficiency. -Translated acquired knowledge to Python, adeptly adapting models to encompass a wide array of Indian Equity Securities, thus expanding and diversifying skill set.

Relevant Coursework: Statistical Learning and Applications, Data Mining, Machine Learning, Natural Language Processing, Data Structures & Algorithms, Information Visualisation, Data Center Scale and Computing, Advanced Computer Vision, Linguistic Modelling, Project Management.