Devansh Sood

Project Details-

• Led a team in the creation of templates with an integrated document editor, streamlining the contract drafting process.
• Spearheaded the development and implementation of workflow-based negotiation solutions, enhancing efficiency and saving valuable time for business teams.
• Designed and oversaw the development of an intuitive platform tailored for use by legal and business teams, emphasizing usability and accessibility.
• Consolidated multiple functionalities into a single platform, eliminating the need for separate tools for drafting, editing, negotiation, search, storage, analytics, and e-signature.
• Used Python Django for scalable backend, React for seamless frontend, MySQL for data management, Kafka for real-time processing.
• Implemented Random Forest using PyTorch for name similarity
• Implemented LSTM using TensorFlow for gender prediction
• Tech Stack - Python Django, MySQL, Rest API, React.

• Worked on the development of a Face Matching Machine Learning project, integrating advanced technologies for facial recognition, similarity analysis, and liveliness detection.
• Integrated features to enhance user security, including liveliness detection methods such as facial gestures and one-time passwords (OTPs).
• Implemented text recognition to extract OTPs from images, ensuring a robust multi-factor authentication system.
• Utilized FastAPI for the back-end to establish seamless communication, providing a high-performance and asynchronous interface for image input, result retrieval, and multi-factor authentication.
• Employed a recognition service to address challenges like variations in lighting, facial expressions, and angles.
• Played a key role in designing a user-friendly interface for image input, result retrieval, and multi-factor authentication.

Project Details-

• Worked on the development of a Face Matching Machine Learning project, integrating advanced technologies for facial recognition, similarity analysis, and liveliness detection.
• Integrated features to enhance user security, including liveliness detection methods such as facial gestures and one-time passwords (OTPs).
• Implemented text recognition to extract OTPs from images, ensuring a robust multi-factor authentication system.
• Utilized FastAPI for the back-end to establish seamless communication, providing a high-performance and asynchronous interface for image input, result retrieval, and multi-factor authentication.
• Employed a recognition service to address challenges like variations in lighting, facial expressions, and angles.
• Played a key role in designing a user-friendly interface for image input, result retrieval, and multi-factor authentication.
• Tech Stack - Machine Learning, Python, Fast API, Web Socket, Open CV, NLP, LLM, OCR, AWS Rekognition

Project Details-

• Worked on Machine Learning based projects specializing in identifying the similarity between two names.
• Successfully addressed challenges including typos, missing characters, missing name parts, jumbled names, different last/first names, phonetic similarity/disimilarity, and varying genders.
• Achieved a remarkable 94% accuracy rate in identifying name similarity, surpassing the industry standard.
• Significantly reduced the false prediction rate from 85%, by implementing advanced machine learning models.
• Integrated Kafka for continuous logging, ensuring real-time tracking of predicted results and enhancing system monitoring capabilities.
• Utilized React.js for the front-end and Flask for the back-end to establish seamless communication, providing a responsive and user-friendly interface for data input and retrieval.
• Currently, model Trained is able to recognize the above mention use cases with an accuracy of 94%.
• Tech Stack - ReactJs, JavaScript, Python, Flask, Kafka, RESTful APIs, TensorFlow, PyTorch, AWS Sage maker, MongoDB, Machine Learning, NLP, LLM.

Project Details-

• Developed a microservice for the Stamp Panel dashboard, which served as a centralized platform for various teams involved in procuring stamp papers. The microservice provided a single-place solution for streamlining the entire process.
• Implemented a user-friendly Cart Facility feature that allowed users to easily add and remove stamp papers that needed to be ordered. This feature simplified the selection and management of stamp papers, improving the efficiency of the ordering process.
• Integrated a payment and invoice generation system to facilitate smooth transactions with vendors.
• Incorporated a comprehensive recording and tracking system to keep track of all courier-related information. This functionality provided real-time updates on the status and location of stamp papers during transit, ensuring transparency and efficient delivery.
• Leveraged Python, Django, RESTful APIs, React, machine learning/OCR, and Bash scripting to build the microservice. These technologies enabled seamless communication between different components, enhanced user experience, automated document processing through OCR, and facilitated system administration tasks through Bash scripting.
• Tech Stack- MySQL, Python, Django, RESTful APIs, React, machine learning/OCR, and Bash scripting, NLP, LLM.

Project Details-

• Created a Django and React-based dashboard for strategy monitors to view and handle errors associated with running strategies.
• Developed an Error Management System within the dashboard, reducing the reliance on external platforms like Slack for error notifications.
• Contributed to the Emergency Response Team (ERT) to address critical situations and resolve urgent issues related to strategy execution.
• Implemented an approval system as a security measure, allowing traders to raise multi-level tickets for changes in the parent strategy.
• Utilized Python, React, Django, Bash/Shell scripting, and RESTful APIs in the development process.
• Tech Stack - Python, React, Django, Bash/Shell scripting, and RESTful APIs

Project Details-

• Worked on a project to track daily stock buying and selling activities, involving multiple reconciliations.
• Performed manual reconciliations as well as automated them using scripts for increased efficiency.
• Utilized Django, Python, and Bash scripting to enhance the functionality of the trading dashboard.
• Automated day-to-day tasks to improve business efficiency and streamline operations.
• Successfully handled reconciliation issues for multiple accounts, ensuring accurate tracking of stock transactions.
• Optimized existing scripts and automated remaining reconciliation processes for improved efficiency.
• Leveraged Python and Bash/Shell scripting to accomplish these tasks.
• Tech Stack - Django, Python, and Bash scripting

Project Details-

• Store Demand Forecast was an advanced version of the Long Range Forecast.
• Focused on dynamically predicting the stock requirements for each item in various stores by leveraging historical sales and discount data.
• Utilized Python and Bash/Shell scripting to further enhance the functionality and accuracy of the forecasting model.
• Collaborated with the team to refine and improve the existing forecasting algorithms and methodologies.
• Tech Stack - Python and Bash/Shell scripting

Project Details-

• Contributed to a project for a UK-based company involving sales prediction for their stores.
• Utilized pre-provided business logic and historical sales data from the previous year to predict store sales for the next 5 years.
• Implemented the prediction models using Python and Spark, leveraging their powerful data processing.
• Developed algorithms and coding solutions to accurately forecast future sales and estimate store demand for specific products based on the client's requirements.
• Employed Python for data analysis, feature engineering, while utilizing Bash/Shell scripting for automation and data processing tasks.
• Tech Stack - Python, Bash/Shell scripting