Tim Roller, CFA
AI Engineer · 15 years across trading, data, and AI
I build agentic systems that combine real-time data, document retrieval, and web research. Prior life as a derivatives trader means I think natively in time-series data, probabilistic systems, and building under uncertainty.
Featured Project
Finance Research Agent
An AI agent that autonomously researches stocks by combining three data sources into structured research briefs.
Agent SDK
RAG
MCP
Python
Live Demo
finance-agent — MU research brief
$ uv run python -m finance_agent.agent
Finance Research Agent
2026-04-10 18:42
> Research Micron (MU) — combine SEC filings with latest earnings
Calling: get_stock_quote → MU: $420.59 (52w: $65.65 – $471.34)
Calling: search_sec_filings → 5 chunks from MU 10-K (MD&A, Risk Factors)
Spawning: researcher subagent → 3 web searches (earnings, analysts, HBM)
───────────────────────────────────
Micron Technology (MU) — Research Brief
Price: $420.59 | Market Cap: $474B
FQ2 2026 Earnings (Mar 18):
Revenue: $23.86B vs $19.6B est (+22% beat)
EPS: $12.20 vs $8.79 est (+39% beat)
Margin: 74.4% gross (record)
FQ3 Guidance (Apr 9):
Revenue: $33.5B vs $22.4B consensus (+50%)
From SEC 10-K (RAG):
HBM supply sold out through 2026 at fixed prices
Only US-headquartered HBM supplier (CHIPS Act moat)
HBM4 in volume production for NVIDIA Vera Rubin
Bull case: $550-700 (HBM4 leadership, structural AI demand)
Bear case: $300-380 (priced for perfection, Samsung recovery)
Saved to: research/2026-04-10_MU.md
───────────────────────────────────
Agent finished. Tokens — Input: 45,230 | Output: 3,842
Latest Writing
I’ve been exploring how agentic AI changes the way we search financial documents. I built a system that searches 4,575 SEC 10-K filing chunks using three different approaches — naive...
12 Apr 2026
All Projects → · All Writing →
