AI and mainframe modernization for regulated industries
Pozent straddles legacy and cutting-edge: a mainframe shop (COBOL, DB2, IBM Mainframe, Endevor) now layering generative AI infrastructure (LangChain, LlamaIndex, Pinecone, Weaviate, FastAPI). The tech-stack shape — legacy data systems + vector databases + LLM tooling — reveals a company helping regulated enterprises (healthcare, pharma) migrate workloads from on-premises mainframes to cloud while embedding AI-driven analytics and clinical data platforms. Hiring is heavily senior engineering-focused (18 of 22 open roles), concentrated in the US and India, signaling rapid delivery demands on both modernization and new AI product work.
Pozent is a digital transformation and workforce solutions company founded in 2012, headquartered in Piscataway, New Jersey, with 51–200 employees. The company operates across three core domains: mainframe systems modernization (COBOL, DB2, Endevor, ISPF), cloud infrastructure (AWS, Azure, GCP), and AI applications (generative AI, machine learning, retrieval-augmented generation pipelines). Their customer base skews toward regulated industries, particularly healthcare and clinical operations, where compliance with FDA, EMA, and ICH-GCP standards is non-negotiable. Revenue comes from both custom development and staffing solutions; the latter feeds a talent pipeline for client IT departments.
Pozent combines legacy enterprise (COBOL, DB2, Mainframe, Endevor) with cloud (AWS, Azure, GCP) and AI tools (LangChain, LlamaIndex, Pinecone, Weaviate, Python, FastAPI). They also use Microsoft 365 (Teams, SharePoint, Power Apps, Power Automate) and vector databases for generative AI pipelines.
Current projects include generative AI solutions with LLM fine-tuning and RAG pipelines, clinical operations data platforms, enterprise analytics, CI/CD for ML workflows, AI-driven APIs, and agentic AI systems with autonomous decision-making. A significant focus is healthcare compliance and regulated data handling.
Other companies in the same industry, closest in size