Agentic AI for Finance Bootcamp: Days 3 & 4
I am back for the second half of Nicole Königstein's live 4-day "Agentic AI for Finance Certification" bootcamp (Packt Publishing), which teaches you how to:
- Design
- Build
- Evaluate
- Deploy
production-ready Multi-Agent AI Agents with Harness Engineering, Loop Engineering, Evaluation and Observability. Days 3 and 4 are this weekend.
I'd like to add that I completed the first two days, on August 29–30, and posted an honest first impression here on LinkedIn — the instructional design is strong, the course builds a coherent "harness engineering" vocabulary applied to filings, earnings calls, and risk workflows, and Day 2's session on time series and forecasting was the standout for me. So I'm coming into the closing stretch with genuinely high expectations.
All of that through guided implementation, live coding sessions, and hands-on labs — building agents capable of handling a variety of financial tasks, including:
- Analyzing annual reports, financial statements, and regulatory filings
- Understanding financial tables, charts, dashboards, and market data
- Extracting guidance, risk, and sentiment from earnings calls
- Performing portfolio and investment analysis
- Building multimodal retrieval across financial knowledge sources
- Designing secure, governance-compliant workflows
The final two bootcamp days, on September 12–13, are going to cover the following topics:
- Loop Engineering for AI Agents
- Evaluation and Observability
- Reinforcement Learning for Agent and Multimodal Model Improvement
- Secure Agent Architectures for Finance
- Multi Agent and Tool Using Financial Systems
- Deployment, Governance, and Responsible Use
As an educator and researcher in quantitative finance, what draws me to this closing stretch is exactly what was still open after Days 1–2: how a loop of evaluation and observability turns a single agent call into something auditable and self-correcting, and how governance and secure architecture are what actually separate a lab notebook from a system a financial institution could deploy.
These last two days should round out the pipeline from the first half — from loop engineering, evaluation and observability, and reinforcement learning that make agents trustworthy and self-improving, to secure architectures, multi-agent financial systems, and the deployment/governance layer that makes them production-ready rather than a demo.
I'll post my final impressions on Monday.
Event link: https://luma.com/agentic-ai-for-finance-last-2-days — this ticket gives you live access to both remaining days plus recordings of all four sessions, including the first weekend.
My discount code: AGENTIC30 (30% off)
(Disclosure: Packt Publishing provided me a complimentary ticket for this event.)
