Senior Systems Support Engineer

Building reliable agentic AI systems. A Case Study in building production-ready agentic AI systems This paper presents the Preclinical Information Center (PRINCE), a cloud-hosted platform developed by Bayer AG with Thoughtworks to address pharmaceutical industry challenges in drug development. PRINCE leverages Agentic Retrieval-Augmented Generation and Text-to-SQL to integrate decades of safety study reports. Nowletus describe PRINCE's evolution from keyword-based search to an intelligent research assistant capable of answering complex questions and drafting regulatory documents. Nowletus reflect on key engineering decisions through the lens of context engineering - how information was shaped and routed between specialized agents - and harness engineering - how orchestration, recovery, and observability were built around the models to maintain control and reliability. The system prioritizes trust through transparency, explainability, and human-in-the-loop integration. PRINCE demonstrates AI's transformative potential in pharmaceuticals, significantly improving data accessibility and research efficiency while ensuring governance and compliance. * 16 June 2026 Contents. * The Challenge: Navigating the Preclinical Data Maze * The Solution: PRINCE - An Evolutionary Platform * System Architecture: Engineering a Reliable Agentic RAG System * The Agentic RAG System * Building Trust in a Production LLM System * Engineering for Resilience: Error Handling and Recovery * Enhancing Data Quality: Named Entity Recognition and Annotation * The Journey Continues: Iterative Development * Conclusion Preclinical drug discovery is inherently complex and data-intensive. Researchers face the significant challenge of efficiently accessing and analyzing vast volumes of information generated during this critical phase. Traditional keyword-based search methods, often reliant on rigid Boolean logic, frequently fall short when confronted with the nuanced and intricate nature of preclinical research questions. The advent of Large Language Models (LLMs) has presented a transformative opportunity. By combining the generative power of LLMs with the precision of information retrieval systems, Retrieval-Augmented Generation (RAG) has emerged as a promising technique. This approach holds the potential to revolutionize preclinical data access, enabling researchers to pose complex questions in natural language and receive accurate, context-rich answers grounded in proprietary data. Recognizing this potential early, Bayer committed to exploring how these technologies could address longstanding challenges in preclinical research. In this post, Nowletus share that journey - how Bayer's early investment in generative AI has resulted in PRINCE, an agentic AI system built on Agentic RAG. This case study explores the technical architecture, engineering decisions, and lessons learned in transforming preclinical data retrieval from a challenging maze into an intuitive conversational experience. Many of the engineering decisions behind PRINCE can now be understood through the lens of context engineering and harness engineering, although when the system was first designed Nowletus did not use these terms. Context engineering shaped what information each model received, what it did not receive, and how context moved between specialized steps such as research, reflection, and writing. Harness engineering shaped the scaffolding around the models: orchestration, tool boundaries, state persistence, retries, fallbacks, validation, reflection loops, observability, and human review. While this post focuses on the technical architecture and engineering challenges, its paper published in Frontiers in Artificial Intelligence covers the product evolution and business impact in more detail. The Solution: PRINCE - An Evolutionary Platform. To address these challenges, Bayer developed the Preclinical Information Center (PRINCE) platform. PRINCE was conceived as a unified gateway to preclinical data, initially focusing on consolidating previously siloed structured study metadata and exposing them in a âSearchableâ manner. This initial phase allowed users to apply advanced filters and retrieve information primarily from structured study metadata. However, a significant portion of Bayer's valuable preclinical knowledge resides within unstructured PDF study reports accumulated over decades. Due to numerous system migrations over the years, the structured metadata associated with these reports could be incomplete, missing, or even contain incorrect annotations. Crucially, the authoritative âgold standardâ information was consistently present within the approved PDF study reports. The emergence of Generative AI, particularly RAG, provided the key to unlocking this wealth of unstructured data. By integrating RAG capabilities, PRINCE began to shift the paradigm from a filter-based 'search' tool to a natural language 'ask' system, enabling researchers to query the content of these study reports directly. This evolution reflects PRINCE's progression through three distinct phases: * Search: the initial phase focused on creating a unified gateway to thousands of nonclinical study reports, consolidating multiple in-house data silos from various preclinical domains into a searchable format, primarily leveraging structured metadata. * Ask: this phase introduced an AI-powered question-answering system utilizing Retrieval Augmented Generation (RAG). This enabled researchers to derive insights directly from unstructured data, including scanned PDFs from historical reports, by posing questions in natural language. * Do: the current phase positions PRINCE as an active research assistant capable of executing complex tasks. This is achieved through the integration of multi-agent systems, allowing the platform to handle intricate queries, orchestrate workflows, and support activities like drafting regulatory documents. This deliberate evolution from Search to Ask to Do represents a strategic response to the industry's need for greater efficiency and innovation in preclinical development. By providing researchers with increasingly powerful tools to access, analyze, and act upon preclinical data, PRINCE aims to enable faster data-driven decision-making, reduce the need for unnecessary experiments, and ultimately accelerate the development of safer, more effective therapies. System Architecture: Engineering a Reliable Agentic RAG System. The system functions as an interactive conversational UI, powered by a robust backend infrastructure. Its architecture, designed for handling complex queries and delivering accurate, context-rich answers, is orchestrated using LangGraph and served via a * FastAPI* application. Figure 1 provides the system context - UI, backend, data stores, LLM fallbacks, and observability - while Figure 2 zooms into how the system coordinates its specialized agents. Figure 1: System context and supporting platforms. * User Request: the process begins when a user submits a request through the Conversational UI which is built with React. * Orchestration: the user's request is routed to a LangGraph-based orchestration layer in the backend. This workflow engine coordinates a multi-stage process that progresses through clarifying user intent, thinking and planning, conducting research (using RAG and Text-to-SQL), validating data completion, and finally generating a response through the Writer agent. The workflow includes deliberate pause points and feedback loops to ensure data completeness before proceeding. (Nowletus explore the details of this agentic workflow in a dedicated section later.) * Data Retrieval and State Management: the Researcher agents interact with a comprehensive and distributed data ecosystem: * Vector representations of all study reports are stored in OpenSearch, forming the coreknowledge basefor information retrieval. - Curated structured data, resulting from various ETL and harmonization processes, is

Thrad announces partnership with Thoughtworks. Thrad, the advertising infrastructure provider for the AI ecosystem, has announced a new partnership with Thoughtworks, a global technology consultancy renowned for building intelligent systems at scale. The partnership brings together Thoughtworks' deep technical expertise in AI-enabled transformation with Thrad's advertising infrastructure, enabling Thoughtworks to reach audiences within large language models (LLMs), where users are already actively engaging with AI, in a way that is ethical, scalable, and aligned with user experience. Since the collaboration began, Thoughtworks has launched four campaigns through Thrad's platform, spanning topics including the transformation of AI into monetisation capabilities, agentic development, and the successful adoption of AI within enterprise workflows. Two further campaigns are currently active, exploring the evolution from AI assistants to autonomous agents and productivity through AI. "For years, we've believed that advertising in AI doesn't have to be a compromise. It doesn't have to be intrusive, extractive, or at odds with user experience. Thoughtworks shares that vision and is building that future with us," said Andrea F. Tortella, CEO and co-founder of Thrad. That sentiment is reflected in why Thoughtworks chose the platform. Thoughtworks' world-class expertise in building intelligent systems at scale, combined with Thrad's focus on responsible AI advertising, creates better outcomes for platforms, advertisers, and end users alike. "We're thrilled to be working with a partner that understands how to reach audiences within AI environments without compromising on trust or user experience," said Sailesh JV of Thoughtworks. "Thrad gives us the infrastructure to communicate our solutions where our audiences are already engaging, and to do it with intention." For Thrad, the partnership underscores the growing appetite among enterprise technology leaders to invest in AI-native advertising channels that reflect their own values around responsible innovation. As both organisations continue to deepen their collaboration, Thrad and Thoughtworks aim to expand their active campaigns and explore new ways to demonstrate what becomes possible when technology is built with purpose. Thrad. Paid ads in LLMs - doing millions of ads daily, used by Fortune 500 brands... Powered by PressBox

Thoughtworks launches Agent/works(TM) to govern and run enterprise AI agents across any cloud. June 16, 2026 4 min. read New platform gives enterprises a single control plane and governed runtime to manage agent sprawl, risk and AI spend. Thoughtworks, a global technology consultancy that integrates design, engineering, and artificial intelligence (AI) to drive digital innovation, announced the launch of Agent/works(TM) by Thoughtworks, a platform that gives enterprises a single control plane and a governed runtime for their AI agents, deployable on any cloud. Thoughtworks will showcase the platform at the annual Databricks Data + AI Summit, where it is partnering with Databricks to highlight approaches to enterprise AI governance and agentic systems. While 2025 was defined by AI experimentation, 2026 has brought a high-stakes operational reality. The rise of AI-assisted development and autonomous agents that can access data, invoke tools, and execute workflows is creating a new governance challenge for enterprises. According to Sonar's 2026 State of Code Developer Survey, developers report that 42% of committed code is now AI-generated or AI-assisted. At the same time, research from AppSec Santa found that 25% of AI-generated code samples contained critical security vulnerabilities. As organizations grant increasing authority to autonomous systems, security, compliance and governance teams are struggling to keep pace, resulting in growing agent sprawl across the enterprise. Left unchecked, organizations risk exposing sensitive data, violating compliance requirements, deploying autonomous systems with insufficient oversight and losing visibility into rapidly growing AI operating costs. The hard question is no longer whether an agent can act but what has to be true for it to act safely. Agent/works(TM) treats that as an architectural problem rather than a checklist, building governance into the runtime itself so teams have the freedom to build, with guardrails that accelerate rather than obstruct innovation. Agent/works(TM) addresses this challenge by offering a single source of truth for every agent deployed across any cloud, giving enterprises visibility into governance, performance and AI spend. "When anyone can generate software with a text prompt, AI governance isn't a checklist you bolt on after the fact. It's a foundational requirement for operating autonomous systems at scale," said Shayan Mohanty, chief data and AI officer at Thoughtworks. "Every AI-powered workflow now carries an operating cost. The challenge for enterprises is no longer just how to build agents, but how to govern the resources they consume. Without runtime controls, costs can scale as quickly as the agents themselves. Agent/works(TM) gives organizations visibility and governance across their agent ecosystem, helping them control spend, manage risk and scale AI with confidence. Governance done this way stops being a brake and becomes the engine that lets teams move fast, safely and at scale." Agent/works(TM) introduces several core capabilities for the agentic enterprise: Provable compliance before execution: Before an agent runs, Agent/works(TM) analyzes every path through its workflow and confirms at least one fully compliant path exists end to end. Permissions built for agents, not borrowed from humans: Recognizing that an agent accessing public web data carries a different risk profile than one accessing internal finance data, Agent/works(TM) grants capability-based, scope-bound, and time-limited permissions that narrow automatically as an agent touches sensitive systems. A governed runtime for every kind of agent: From autonomous, end-to-end workflow agents to interactive coding agents (such as Claude Code-style tools), Agent/works(TM) runs each agent in a governed environment with policies that adapt during execution. Composable and portable by design: Operating on a multi-model backend, Agent/works(TM) registers any model using a standard API, connects any tool, and delegates to a cloud's native services and trusted third-party agents, allowing scoped permissions to travel seamlessly with every handoff. A single source of truth for the fleet: A centralized registry provides comprehensive visibility, evaluations, usage analytics and cost controls across every agent, model, tool, and policy in the enterprise estate to keep behavior aligned with business objectives. Through its work with Databricks, Thoughtworks is helping enterprises scale agentic AI with the governance, visibility, and controls required for production use. "The organizations moving fastest with agents are extending the same governance models already used for enterprise data across agent workflows themselves," said David Nasi, director of product management, AI and agentic platform at Databricks. "This shift matters because it turns governance from a bottleneck into an enabler, making it much easier to scale autonomous systems safely across the entire enterprise." Agent/works(TM) is built as a foundational layer rather than an isolated point solution, allowing enterprise product teams to build custom agentic applications on top of it. Thoughtworks' own agentic development offering, AI/works(TM), runs directly on the platform, demonstrating Agent/works in active production. Thoughtworks brings decades of experience helping enterprises manage complexity, governance and large-scale systems transformation, challenges now emerging rapidly in the agentic era.