Skip to content

Shubhankar Kahali

RSS Feed

I began my career as a security researcher and later transitioned into tech recruiting—an uncommon path that gives me a distinct advantage. I stay close to the field through malware analysis and by participating in bug bounty programs with companies like Google, Apple, Meta, Microsoft, Amazon, and Uber. This background gives me insight into the real patterns of technical thinking—what separates great engineers from good ones. I write about the intersection of technology and human potential.


Today my work splits across technical recruiting and machine learning research. On the recruiting side, I design evaluations that go beyond interviews and I headhunt talent for unique, niche roles. On the research side, I'm focused on sequence modeling for large-scale foundation models, and on challenges that emerge when models meet the messiness of real-world deployment. I'm also exploring how the application of foundation models will reshape computing interfaces over the coming decades—especially for deep, exploratory work like data science. Concretely, I work with neural architectures, attention mechanisms, and few-shot learning, from cognitive systems to transformer optimization, with applications in aerospace, defense, and cybersecurity.


I'm also building Glide Logo, my weekend project implementing transformer-based multi-agent systems for career intelligence. The architecture leverages Kubernetes microservices, vector similarity search, and Kafka streaming for real-time feature engineering, with core ML pipelines processing 200+ attributes through gradient boosting ensembles and neural collaborative filtering. The system currently serves 39K+ users with 94% job match accuracy, delivering $21.9M+ in salary optimization and 5,605+ successful placements—validating production-scale ML system design.


Recent Posts

All Posts