Laya Myadam

Full-stack AI/ML Engineer — React/TypeScript · Python FastAPI · LLM APIs (GPT-4, Claude, Gemini) · Multi-Agent Systems · RAG Pipelines · RL · Graph ML · AWS — building intelligent applications across health, finance, and education domains Built production-grade multi-agent systems (LangChain, LangGraph, AutoGen) with RAG, MCP, and LLM routing — achieving 87%+ accuracy and 40% reduction in human intervention for 1,000+ users Developed LLM evaluation frameworks with safety checks, hallucination detection, and real-time dashboards monitoring 15+ metrics — reducing model failures by 92% Fine-tuned LLMs (DistilBERT, T5, LLaMA) with LoRA/QLoRA, vector DBs (FAISS, Pinecone), RL agents (PPO, Bandits), and GAT graph models — boosting F1 by 18%, engagement by 16%. Deployed on AWS (ECS, SageMaker, Lambda) with Docker + CI/CD — cut deployment from 3 weeks → 2 days, sub-200ms latency, 99.5% uptime

Guided students on low-level computing concepts (memory, CPU execution, instruction pipelines), strengthening understanding of performance-critical systems. Conducted debugging sessions and code reviews, helping students optimize programs for efficiency and correctness. Applied analytical problem-solving and algorithmic thinking to break down complex programming challenges. Supported students in understanding how hardware-level execution impacts software performance, relevant to ML model optimization and inference efficiency.

Developed daily sales dashboards using Power BI, Excel, and MS Access to support financial reporting and operational decision-making. Managed procurement, requisitions, tax reallocations, and stored value transfers while ensuring accuracy in budget-related records. Maintained budget-spent data in MS Access and supported the Director of Budget & Operations in budget preparation, reporting, and analysis. Performed data cleaning, transformation, and validation using Python, pandas, and NumPy to improve reporting quality and process efficiency. Applied basic ML/statistical methods including forecasting, trend analysis, and simple predictive models to identify spending patterns and support planning. Supervised student assistants by assigning, monitoring, and reviewing daily operational tasks.

Developed and fine-tuned machine learning models using scikit-learn, XGBoost, and Random Forest for regression and classification tasks, achieving 18–22% reduction in RMSE through advanced feature engineering and hyperparameter optimization. Built deep learning architectures with PyTorch, including CNNs for image recognition and LSTM/GRU networks for sequential text data, boosting model accuracy by 15% and lowering inference time by ~20%. Implemented training pipelines for predictive maintenance applications, leveraging distributed computing and parallel processing to reduce training durations by 40% and detect failures proactively. Designed end-to-end NLP pipelines using TF-IDF, Word2Vec, GloVe embeddings, and early BERT fine-tuning, deploying scalable REST APIs with FastAPI, Docker, and AWS, improving F1-scores and cutting inference latency by 25–30%.