Calvin Ruben Raj C

AI/ML Engineer,

• Worked for one year on training, fine-tuning, and evaluating locally deployed Large Language Models (LLMs) using Python.
• Prepared and cleaned training datasets through data collection, preprocessing, tokenization, deduplication, and quality validation.
• Fine-tuned open-source language models on domain-specific datasets using PyTorch, Hugging Face Transformers, and parameter-efficient techniques such as LoRA/QLoRA.
• Developed Python scripts and automated pipelines for model training, evaluation, inference, and experiment tracking.
• Optimized models for local deployment using quantization, prompt optimization, and memory-efficient inference techniques.
• Evaluated model performance using accuracy, perplexity, response quality, latency, and task-specific metrics.
• Built local LLM applications involving prompt engineering, semantic search, embeddings, and Retrieval-Augmented Generation (RAG).
• Diagnosed training and inference issues, including overfitting, poor response quality, GPU memory limitations, and slow execution.
• Documented experiments, model configurations, evaluation results, and deployment procedures.
• Collaborated with team members to improve dataset quality, model reliability, and application performance.

Technologies: Python, PyTorch, Hugging Face Transformers, PEFT, LoRA/QLoRA, LangChain/LlamaIndex, FAISS/ChromaDB, Pandas, NumPy, Git, Jupyter Notebook, CUDA, Ollama/llama.cpp.

NMSWorks has two decades of experience in delivering OSS/BSS solutions for the end-to-end management of carrier grade networks.

• Supported product managers and development teams through all testing phases, ensuring thorough planning, execution, troubleshooting, and reporting in a high-volume agile environment.
• Initiated automation testing which reduced manual testing and time consumed we are able to deliver software faster.
• Development of automated testing using script(Bash and Python).
• Performed Functional Testing with multiple user scenarios,
• Test analyses using log, JFR and Jstack(Backend Testing).
• Product that i have tested in company: CygNet-NG, CygNet-VIMS, CygNet-IAMS, Report Portal, Router Analytical.
• Deployment of build and Server setup(apache).
• Executed data migration to integrate initial datasets into the application.
• Performed regular maintenance tasks on Oracle database to ensure optimal performance.

Company delivers VR based stimulation Products. Animation for Advertisement worker as freelancer.

• Import and Export data in mysql data base using django
• Work in agile methodology.
• Use to send data from backend to Android as rest API using django framework.

Chinese company, Creates the Closed caption for animation series and movie and English tutorial videos

• Typed captions with accurate pronunciation and punctuation for various media.
• Delivered high-quality subtitles for multiple projects, ensuring clarity and consistency.