Designed and developed internal AI agents using Large Language Models (LLMs) and python to automate business workflows and data processing tasks. Developed RAG-based AI agents using LangChain and Python. Integrated LLM-powered solutions with internal APIs and data sources. Implemented prompt engineering strategies. Optimized performance and latency of LLM applications. Deployed and tested AI solutions using AWS Bedrock.

Perform full-trace labeling to train AI systems using Test-Driven Development (TDD) methodology. Work inside AWS SageMaker environments and AWS-managed servers. Use Docker for isolated development. Implement TDD workflows with pytest and Python to debug and validate code. Document problem-solving steps using DataGenToolKit.

Integrated LLMs to map MCC codes. Built merchant logo retrieval via BeautifulSoup and GANs. Upgraded spaCy and implemented GPU-accelerated training. Developed a TensorFlow LSTM model for salary detection. Led location extraction study using OpenAI, LLaMA, and DeepSeek. Implemented data retrieval pipelines for RAG workflows. Designed merchant name deduplication with FuzzyWuzzy and FAISS. Created an LLM-based keyword extractor.

Proposed a business identification model product. Integrated API queries and AWS S3 file consumption. Developed NLP models in Python for transaction classification and named entity recognition (NER) using spaCy. Deployed and maintained ML models on AWS SageMaker. Conducted experiments with transformers and LLMs. Conducted benchmarking for production NLP applications.

Managed customer database connections to Redshift via Fivetran. Monitored transaction volume and ML model execution via AWS Lambda. Created dynamic reports with Streamlit, deployed using GitHub Actions and AWS ECR.

Maintained and deployed transaction classifier models using AWS services. Developed clustering models and optimized classifier code. Monitored services for data consumption.

Developed TensorFlow LSTM models for time-series forecasting in mining. Designed and trained Computer Vision models for occupational health and safety. Built deep learning solutions for environmental monitoring. Researched image classification using CNNs. Performed end-to-end ML development from data preprocessing to deployment.

Consultancy for scientific research and thesis projects. Developed ML models with Python and TensorFlow. Applied process mining techniques.

Enhanced maintenance programs for ISO certifications. Managed chemical dosages and lab process analysis. Conducted statistical analysis with R and Python. Updated production costs and performed time series analysis.