Open to opportunities

rajesh datta sanka

@rajeshdattasanka

Message

Machine Learning Engineer specializing in scalable AI, LLMs, MLOps, and privacy-aware systems.

United States

Message

What I'm looking for

I'm seeking a hands-on ML/MLOps role building production-grade LLMs and real-time systems, emphasizing privacy, explainability, collaboration, and measurable business impact.

I am a Machine Learning Engineer with 4+ years of experience designing end-to-end ML systems, deploying production models, and driving applied AI research.

I have built NLP and RAG pipelines, real-time recommendation and credit-risk systems, and graph-based fraud detection models that materially improved business metrics (intent recognition, response relevance, false-positive reduction, latency, and model refresh cycles). I routinely deploy services with Docker, Kubernetes, AWS and GCP, automate retraining with MLflow and CI/CD, and apply explainability and differential privacy techniques.

I work across NLP, computer vision, anomaly detection, and forecasting with a strong focus on human-centered AI, interpretability, and secure modeling. I seek roles where I can scale production ML, mentor teams, and deliver measurable impact.

Experience

Work history, roles, and key accomplishments

Current

Machine Learning Engineer

Current

Capital One

Jun 2022 - Present (3 years 2 months)

Built NLP and RAG pipelines for the virtual assistant and internal search, increasing intent recognition by 19% across 12M+ daily queries and improving response relevance by ~27%. Deployed LLM microservices and automated retraining workflows, reducing latency from ~1.8s to 900ms and improving model refresh cycles by 40%.

Python PyTorch Transformers AWS Kubernetes fastAPI OpenSearch MLFlow TensorFlow

Machine Learning Engineer

HCL Technologies

Jan 2020 - May 2022 (2 years 4 months)

Led predictive maintenance and ML solutions that reduced downtime by 30% and saved $4M+ annually, and engineered NLP pipelines improving throughput by 25%. Delivered forecasting and vision models that boosted accuracy by 11% and achieved 95%+ precision on defect detection while building MLOps pipelines to cut deployment timelines by 40%.