How LTM Uses Snowflake Cortex AI to Predict Candidate Onboarding with 80% Accuracy

LTM is a global technology consulting and digital solutions company headquartered in India, operating at the scale where hiring is both a core operational function and a major financial risk. With more than 80,000 users depending on its talent acquisition infrastructure and post-offer dropout rates as high as 30%, the cost of reactive hiring decisions — paying one to three times a candidate’s salary to re-recruit — was a persistent drag on operations during peak hiring seasons.

Before migrating to Snowflake, LTM’s HR analytics team was constrained by legacy, on-premises systems built around data silos. Candidate predictions were delayed by weeks, ETL pipelines required constant maintenance, and prediction models couldn’t scale to match hiring volumes. Engineers spent the bulk of their time searching through disparate data sources rather than building predictive capabilities. As Rajeev Menon, Executive Vice President of Human Resources, described it: “The hiring landscape is evolving rapidly, and technology is no longer just an enabler — it is a strategic partner in shaping better candidate experiences and driving measurable outcomes.”

LTM migrated to Snowflake in deliberate phases: first using Snowconvert for a lift-and-shift of critical workloads, then modernizing ETL pipelines with zero-ETL connectors to SAP, SuccessFactors, and ATS platforms, and finally deploying ML models via Snowpark for end-to-end pipeline operations. Snowflake Cortex AI provides access to LLMs including Arctic, Llama 3, and Mistral through a fully managed serverless environment, enabling the team to build and deploy AI applications five times faster than their previous approach. The joining probability predictor — a core ML model analyzing 12 predictive features from historical joiner and dropout data — segments candidates into Red, Amber, and Green categories based on onboarding probability.

The operational results were significant across multiple dimensions. LTM now predicts candidate onboarding probability at over 80% accuracy, giving hiring teams a 25-to-30-day window to intervene with personalized offers or culture-fit sessions before a potential dropout. Query performance on critical hiring data improved 10x. Total cost of ownership dropped 70% compared to legacy systems. AI and ML application development accelerated 5x. More than 700 users across Talent Acquisition, Business, and Operations access these real-time insights through Power BI dashboards built on top of the Snowflake data layer.

LTM’s trajectory from reactive to predictive hiring is now built on an enterprise-ready foundation. The combination of Cortex AI’s LLM access, Snowpark’s modular processing architecture, and precision governance meeting both GDPR and India’s DPDP Act requirements positions LTM to scale AI safely across its global workforce. The platform is the infrastructure layer for a talent strategy that treats hiring prediction not as a back-office function but as a measurable competitive capability.