Osei Brempong

• Designed and deployed AI-powered conversational solutions, integrating Conversational AI, Agent Assist, and Co-Pilot AI, improving response accuracy by 35% and reducing user query resolution time through optimized AI-driven interactions.
• Led research in NLP, reinforcement learning, and AI-driven automation, increasing chatbot efficiency by 40% and reducing customer handling time by 25%, leveraging techniques such as self-supervised learning, large language models (LLMs), and AI agentic workflows to enhance user engagement.
• Took end-to-end ownership of AI/ML models, managing the entire lifecycle from data collection and labeling to training, inference, deployment, and scaling, ensuring AI solutions are robust and production-ready.
• Collaborated with cross-functional teams in engineering, product, and professional services, reducing AI solution integration time by 30%, improving system scalability, and increasing AI adoption rates by 50%, while maintaining integrity, a strong work ethic, and an innovative, outside-the-box approach to solving AI challenges.

• Engineered Transformer-based architectures (BERT, GPT, T5) for named entity recognition (NER), text classification, and sentiment analysis, achieving state-of-the-art (SOTA) performance on benchmark datasets.
• Optimized sequence-to-sequence (Seq2Seq) architectures for neural machine translation (NMT) and abstractive text summarization, implementing low-rank adaptation (LoRA), and quantization-aware training (QAT) to reduce inference latency by 20%.
• Implemented machine learning algorithms such as random forests, and SVMs for predictive modeling and feature engineering, improving classification performance in real-world datasets.

Developed and maintained scalable web and mobile applications with a focus on high availability, low latency,

and efficient memory management, leveraging React, React Native, and Flutter, ensuring 99.9% uptime and

sub-200ms response times.

Key achievements include improving app responsiveness by 30%, enhancing security with multi-factor

authentication (MFA) to reduce unauthorized access by 90%, and boosting service bookings by 35% with a Quick

Repair Request feature.

Integrated third-party APIs and payment gateways (Stripe, PayPal, Razorpay) with secure OAuth 2.0 & JWT

authentication, ensuring PCI-DSS compliance and reducing transaction failures by 20%

• Implemented robust data quality assurance protocols using SQL constraints, automated schema validation, and cross-region reconciliation scripts, ensuring 99% data consistency across distributed health information systems
• Designed advanced analytics frameworks integrating Bayesian inference and time-series forecasting, impacting 10 or more key healthcare policies, and increasing decision-making speed by 30%.
• Achieved a 35% reduction in query execution times by optimizing SQL queries and deploying effective indexing and partitioning strategies