As an AI Engineer, I build AI applications for legal data analysis.Responsible for:• Building data processing pipelines for effective document indexing• Creating and Integrating vector databases for efficient data mining - search and retrieval• Building an LLM sidekick for knowledge workers in the legal space

At Thirdi, we are offering AI automation services to clients and pivoting into a single product-based company.• Created end-to-end data extraction pipelines by scraping the web to: - analyze the housing properties - extract information about SMEs - identify car sellers/buyers• Implemented POCs for GenAI applications: - Conversational bots - AI Agents

• I am positioned to contribute significantly to our team's efforts in integrating AI models into web and mobile applications. I have collaborated closely with cross-functional teams to ensure seamless workflows(automated data pipelines, model training and evaluation, and CI/CD-GitHub Actions) and deliver high-quality AI-powered solutions.• AI Model Enhancement: Improving the object detection model(YOLOv8) on mobile devices using ONNX, quantization techniques, and an image quality model to enhance image capture precision.• I utilized AWS services Sagemaker to build the object detection model, S3 data lake, and EC2 instances.• Utilized AWS API Gateway and Sagemaker Inference to enhance the scalability and responsiveness of computer vision and SaaS platform applications, integrating AI capabilities into backend solutions.

• Deployed LLM applications on GCP using FastAPI, Cloud Run, Pub/Sub, and Cloud Functions, adopting a modular approach for MLOps with distinct feature, training, and inference pipelines to optimize backend processing and application responsiveness.• Automated data management practices by integrating Crunchbase data into Snowflake, leveraging Google Cloud Functions and Snowpipe, and developing ML applications with Snowpark and Mlflow.• Developed a regression model in Snowflake using Snowpark, which was instrumental in guiding investors' strategic decisions by forecasting company returns. Automated weekly predictions with SQL tasks, offering investors continuous, data-driven insights for enhanced investment strategies. Used Ray from Anyscale to speed up the training phase by parallelizing multiple jobs. Orchestrated the workflow using Prefect to keep track of the processes.• Developed a BERT-based recommendation system using Snowpark and Python, specifically designed to aid investors by providing real-time, personalized company insights. Integrated into a Streamlit web application, this system uses user-defined functions for accurate inference, helping investors identify and capitalize on high-potential investment opportunities.• Built a Q&A bot using LangChain that connected to the Snowflake data warehouse, enabling investors, VPs, and Partners to access insights by converting text to SQL queries. Designed and implemented bot logic and conversation flows.

• As a Graduate Student Course Assistant for the graduate-level Algorithms course, I had the privilege of aiding 50+ students in their academic journey. My responsibilities included facilitating a better understanding of complex algorithms concepts, coordinating meetings with the professor when needed, and offering personalized guidance to students to enhance their learning experience.• Additionally, I contributed to the development of course materials. This involved designing and evaluating assignments and midterm examination papers, ensuring they comprehensively assessed students' understanding of the course content.• Throughout this role, I honed key communication, student engagement, and development skills. In addition, this experience underscored the importance of adapting to individual learning styles and fostering an environment conducive to student success.

Throughout my internship with NolymitAI, I leveraged my AI and web development skills to deliver impactful solutions.Key contributions include:• Designed and implemented a multilingual chatbot tailored explicitly for the restaurant industry. This AI-driven solution resulted in a significant increase in orders, decreased customer service requests and improved overall customer satisfaction.• I have developed a comprehensive web application platform designed for small-scale businesses. By integrating machine learning applications into the forum, I contributed to a boost in sales. I also incorporated Firebase, Node.js, and Explicit Content Detection features, leading to a reduction in inappropriate content and enhancing customer retention.• Enhanced project management efficiency by coordinating weekly meetings, preparing detailed minutes, generating strategic work plans, and providing comprehensive work reports. These efforts led to a 30% reduction in project completion time and increased team productivity by 20%.This hands-on experience further solidified my AI and web development expertise and underscored the transformative power of AI and ML applications in real-world business contexts.

As a Software Developer at Vaayu Drones, I tackled complex problems using machine learning. • I developed models to differentiate between cultivated and uncultivated land in drone images, achieving a 30% reduction in image processing time. This problem involved tackling the irregularities of natural landscapes and variations in imaging conditions.• Additionally, I pioneered the development of object detection algorithms for urban planning. These algorithms had to accurately detect and categorize parking spots and roads from drone images, which was quite challenging. However, my efforts led to a reduction in processing time and the successful categorization of 1,000 parking spots and 500 miles of roadways, significantly enhancing city infrastructure planning.• I also spearheaded a project to detect barren land. My machine learning model processed over 500 high-resolution drone images daily and precisely classified over 10,000 square miles of land, enhancing our environmental assessments.These experiences underscored the importance of leveraging machine learning techniques to handle complex data problems. In addition, they reinforced the necessity of technical expertise and understanding of the problem domain to derive meaningful insights.

My tenure as a Machine Learning Intern at Robofied allowed me to combine my technical expertise with my passion for knowledge sharing.Essential tasks and achievements include:• I have authored comprehensive technical articles on Machine Learning and Deep Learning, drawing upon insights from cutting-edge research papers and technical reports. This exercise contributed to my extensive theoretical understanding of the field and offered readers a well-rounded view of these advanced topics.• I showcased my technical proficiency by implementing Machine Learning concepts using Scikit-learn and TensorFlow frameworks, which were integrated into the content of my articles.• My efforts contributed to website traffic and a boost in reader engagement with the articles. These metrics underscored the effectiveness of my approach in making complex topics accessible and engaging to a broad audience.This experience further solidified my understanding of Machine Learning and Deep Learning and demonstrated the value of sharing technical knowledge with a broader community.