Senior AI/ML Engineer at Salesforce

Designed and deployed AI/ML and NLP solutions, led cross-functional teams, and implemented cloud-native architectures on AWS, GCP, and Azure. Built real-time ML pipelines, improved customer engagement and product personalization, and established monitoring and governance to enable scalable AI across the organization.

Lead AI/ML Engineer at iStream Planet

Led scalable live-stream ML integration for major sports events, applying computer vision and deep learning to address video quality and reliability in real-time streaming. Designed real-time inference systems and ML pipelines for batch and streaming use cases, collaborating with product teams to deploy production-ready solutions.

Senior Lead AI/ML Engineer at Salesforce

Architected enterprise multi-agent AI/ML systems and led ML engineering initiatives for a large-scale AI platform; designed agent orchestration, safety mechanisms, and production-grade ML pipelines including RAG, memory-aware pipelines, and multi-tenant embedding pipelines. Implemented end-to-end ML lifecycle, including retrieval, verification, and code synthesis; built low-latency inference stacks with vLLMs and distributed infrastructure. Implemented guardrails, constrained decoding, structured outputs, policy evaluators, semantic diff validation, and confidence-based routing to improve reliability. Built RAG pipelines with semantic indexing, memory-aware batching, vector compression, document re-ranking, and context-window optimization for enterprise knowledge automation. Designed multi-tenant embedding pipelines with parallel encoding, chunking, normalization, vector lifecycle management, and temporal decay strategies. Integrated vector DBs with caching, memory-aware batching, conc

Senior Machine Learning Engineer at iStream Planet

Built end-to-end ML pipelines for large-scale streaming media, including real-time inference, data ingestion, and multi-tenant embeddings pipelines. Designed end-to-end ML platforms leveraging Kafka, Airflow, Spark, Snowflake, and Delta Lake for ingestion, processing, and storage. Implemented memory-aware and batched processing, monitoring, and retraining workflows to support production-grade ML at scale.

Senior Machine Learning Engineer at ForgeRock

Developed ML models and data pipelines for identity and access management, focusing on authentication flows, anomaly detection, and security-oriented ML features. Built scalable ML infrastructure for inference and deployment, collaborating on platform-level ML capabilities and security considerations.

Senior Machine Learning Engineer at Apple

Built transformer and graph-based models to detect anomalous authentication patterns, identity misuse, compromised accounts, and abnormal access behavior. Designed behavioral embeddings and sequential identity event modeling; developed LLML-driven document classification, structured outputs, and compliance-focused validation. Implemented production ML pipelines with robust monitoring and on-device/offline capabilities.

Principal AI/ML Engineer / Principal Data Scientist at ForgeRock

Led end-to-end AI/ML solutions across product lines, architecting scalable ML pipelines, feature stores, and model governance; enabled rapid experimentation and productionization of AI features.

Senior AI/ML Engineer at iStreamPlanet

Designed and deployed AI-driven content ranking systems using transformer-based models to enhance video recommendations, enabling real-time relevance scoring and personalized content delivery.

Senior AI/ML Engineer at StreamPlanet

Led AI-driven product development and an experimentation platform for A/B testing and multi-armed bandits in live broadcasting, evaluating the impact of AI models on user behavior and business outcomes and achieving a 15% improvement in conversion rates.

Principal AI/ML Engineer | Principal Data Scientist at ForgeRock

Designed and deployed AI-driven content-ranking and anomaly-detection systems; built large-scale data pipelines using Kafka, Airflow, Spark, and Delta Lake to enable seamless data ingestion, feature extraction, and retraining for production ML workloads; implemented continuous model evaluation including drift detection and SHAP explainability to meet operational goals and reliability.