Shyam Prasath Email & Phone Number

Bengaluru, KA, IN

Toronto, Ontario, Canada

Toronto, Ontario, Canada

• Led the Machine Learning AI Software Engineering team of 6 members. And Responsible for the product design and architecture of the ML system.
• Developed and contributed to a Multi-model large language model (LLava-1.6). Employing parametric efficient fine-tuning technique and QLORA for fine-tuning for the domain adoptation.
• Developed and launched the pilot version of the research project (Optimize Hydrogen wastage for plant efficiency) convert into production ready product.
• Build complex FAQ Retrieval Augmented Generation (RAG) system with opensource Llama-2 model and vector database using Llama-index and Lang chain Framework.
• Developed Incident management reporting tool using OpenAI GPT-4 preview model to estimate theIncident type, Injuries and Environmental etc... based on the user incident description.
• Design and implemented scalable automated application for managing image annotation for modelbuilding and achieve 28% throughput.

Chennai, Tamil Nadu, India

• Responsible to take complete ownership and carry out each and every step the end-to-end analytics/ML life-cycle process with deployment, performance monitoring of the pipeline in the production environment.
• To ingest streaming/batched real-time data, from multiple sources – SQL-DB, SAS, Cloud etc. Creation of automated pipeline/workflows to ingest streaming/batched real-time data, based on triggers/actions/events.
• Perform Extract-Transform-Load on Millions of rows and columns containing mixed data types [text, numbers etc.]. CRUD operations, Data augmentations, imputations, derived columns etc. Creation of automated.
• ML BASED OPTIMIZATION FOR PARAMETER PREDICTION IN REFINERY OPERATIONS - to build multiclass regression model to optimize the usage of temperature and hydrogen to refine the crude oil. Implemented online wide & deep.
• Optimize the sucrose production and control the wastage of Bassage using ML Technique. The main goal is to optimize the motor RPM and wastage of baggage from each sugar mill.

Toronto, Ontario, Canada

• Implemented optical character recognition(OCR) for invoice processing using Azure API. Integrated RAG-based LLM system to populate the insights from the invoices.
• Developed Django web framework for Employee Expense Reports to Automate vendor invoice processing with power automation and seamless integration with existing SQL database.
• Performed Enhancing PowerBI reporting through DAX and Dynamic dashboard creation under the guidance of the VP of Finance.
• Conducted training and awareness sessions to educate the Finance team on Python and PowerBI.

Mississauga, Ontario, Canada

• Spearheaded the implementation of Bert Transformer model for the weekly-flyers Analysis and achieved significant efficiency improvement by optimizing the weekly flyers comparison.
• Implementation of PowerBI reporting tool to automate data analysis, resulting in reduction in manual reporting efforts and enabling the team to analyze market share trends with greater accuracy and timeliness.
• Achieved significant efficiency improvements by optimizing the manual reporting routines.
• Developed production reports for promotional optimization model, comparing total market tonnage identified opportunities to increase market share.

Bengaluru, Karnataka

• One of my achievements is, Successfully implementing an Automation script which has reduced the manpower resource by 50%. Some of the projects i have Worked/handled:
• Build Web scraping tool for the “Jaboard” product. Using the tool the user can get the organization details from LinkedIn using user defined criteria’s like organization has more than 500 employess and the yearly.
• END TO END VISUALIZATION PLATFORM using opensource. The platform has unique features like without user interpretation the user will get the actionable insights from the data. Plethora of Database connections such as.
• IMPLEMENT SPEAKER DIARIZATION AND VOICE BIOMETRIC. The objective is to build the model which can verify the speaker voice and face from the video file using Gaussian mixture model and speech recognition features. The.

Bengaluru Area, India

• Collaborated towards the AI video conference in-house product, adding new features like face recognition, transcriptions, Voice identification etc..
• Implementing the readable, reusable, and efficient code and proper documentation of the implemented features.
• Database activity and visualization to point out the key insights

Bangalore

As a Trainee, I worked under the BPCL Retail Analytics Team. Providing actionable insights and KPI to make decisions. Assisted the sales team in increasing sales of retail outlooks.Collected the campaign data, to derive the hidden insights of the sales and create ad-hoc reports.Skills: Python, Excel, Power BI

Post Graduate Ontario Certificate, Artificial Intelligence And Data Science

Bachelor Of Engineering & Technology