Vignesh Dharmalingam

Senior AI Architect - Agent Specialist at Arm Cambridge

Led design and deployment of a multi-agent AI system to automate credit analysis for 1,500+ clients; managed a team of five engineers delivering AI products on schedule. Engineered a specialized agent for real-time news analysis of unlisted companies, enabling improved credit management and risk assessment. Validated agent performance on production-grade platforms (Databricks Mosaic AI, Vertex AI). Developed a lightweight agent orchestration framework in Python with custom memory management and established model governance to monitor accuracy and reliability. Served as the interface between business and technical teams to finalize architecture and development plans; advanced LLM performance monitoring and DevOps-driven metrics.

AI Solutions Engineer at British Telecom, Jaguar LR & SKY Telecom

Designed and built retrieval-augmented generation (RAG) pipelines, including chunking optimization, embedding generation, vector DB management (Pinecone), and query processing. Developed prompting techniques to improve outputs, achieving ~15% reduction in operational costs. Automated knowledge extraction with NotebookLM; built data ingestion pipelines using AWS SageMaker; implemented data migration to AWS cloud data warehousing; deployed multi-agent architectures and CI/CD via Azure DevOps.

Lead Data Science Engineer at HCL Technologies

Developed a yoga pose detection pipeline (45 poses) deployed on GCP with Flask; built CNN-based image analysis; created App Engine app, Streamlit frontend, and Cloud Function automation. Contributed to banking data migration (Malaysia); built DataFlow pipelines from Pub/Sub to BigQuery; implemented cybersecurity analytics using PySpark.

Statistical Analyst - Data at Fiserv India

Engineered classification models to predict probability scores in card analytics; developed UI prototypes using Streamlit and Flask; advised on cloud-based solutions on GCP/Azure and cost-effective configurations; collaborated with stakeholders to align cloud architecture.

Software Engineer - Data Science at JK Technosoft

Developed analytics platform based on Kinesis, EMR, Elasticsearch integrated with SageMaker; built NLP-driven recommendation systems; translated customer needs into cloud solutions on AWS/GCP.

R&D Consultant - Data Science at Software AG

Deployed a Google Cloud-based chatbot using Dialogflow; addressed record matching with word2vec and cosine similarity.

Engineer - Data Science at Ericsson (Bangalore)

Analyzed data needs to model telecom functionalities (alarms, downtime, traffic); performed feature selection to reduce dimensionality and improve model performance.