Ghaith Khanfar

Lead Harri's AI Engineering across GenAI/Agentic AI, building the agentic layer on the workforce management and HCM platform, and Data Science & ML. I set technical direction and delivery for both teams and stay hands-on in architecture and code.

• Manage the Data Science & ML team and give technical supervision and support on Harri's AI Demand Forecasting service, which forecasts sales and other metrics that customers rely on for labor deployment and scheduling.
• Lead the Agentic AI team building Salli.

Led the 9-engineer GenAI team that architects and ships Salli, Harri's agentic AI product on its workforce management and HCM platform.

• Took Salli from prototype to General Availability in April 2026.
• Led the Agentic Actions program, moving Salli from read-only Q&A to 80 executable workforce-management actions, each gated behind human-in-the-loop confirmation and covered by a dedicated regression suite.
• Architected Salli's multi-agent system on LangGraph: per-agent ReAct tool-calling loops, an orchestrator with fan-out routing, per-agent model selection across multiple LLMs, token streaming and a surfaced reasoning stream.
• Designed and integrated an A2A/A2UI SDK for agent-to-agent coordination and agent-driven UI, plus Langfuse-based prompt management, tracing and per-user usage analytics.
• Built an internal platform for building agents: a repeatable workflow from architecture discovery and a human validation gate to scaffolding of agents, tools, graph nodes and tests, ending in a spec and PR.
• Shipped 10+ production agents across scheduling and labor, timekeeping, talent acquisition, CoreHR, compliance, local-context awareness and multi-location management.

Owned technical direction and delivery for forecasting, NLP and AI-powered dashboard projects, from requirements and architecture through production rollout.

• Built ML solutions for demand prediction and time-series forecasting, analytics and automation.
• Stack: Python, TensorFlow, Flask, Streamlit, Azure AI.

Led a team of 3 data scientists across 3 client engagements in healthcare and social media.

• Built predictive models for early diagnosis of endometriosis for a femtech client, working with US healthcare data, ICD diagnosis codes and OHDSI ATLAS cohorts. The models were used in the client's research publication.
• Applied NLP and embedding-based language models to topic modelling and clustering of large unstructured text datasets.

Owned the full lifecycle, from data wrangling and feature engineering to training, validation and deployment, of 3 productionized ML services for Harri's workforce management customers. Part-time from Nov 2021, alongside datablock.

• Built predictive models with classical ML and deep learning on structured and unstructured data.
• Built MLOps pipelines automating data processing and model deployment.
• R&D on personalizing ML-powered services for the hospitality vertical.
• Stack: AWS SageMaker, Docker, Python, TensorFlow.

Developed an ML-based resume parser integrated into Harri's hiring and screening workflows.