Writing
Long-form essays on research strategy, applied machine learning, and leadership. Opinionated and specific, written from the perspective of a Senior Applied Scientist positioning for a research leadership role. All essays are original content.
Research agenda
Causal identification and mechanism design for agent-driven decision systems
Three research programs are converging at the intersection of causal inference, LLM-driven decision agents, and mechanism design under uncertainty. The locked thesis, expanded with a specific 18-month research roadmap and five-year vision. Read first if you are evaluating my research direction.
The three causal ML research frontiers that actually matter in 2026
An opinionated survey of where causal-ML research should direct effort over the next three years: panel-data methodology with heterogeneous treatment effects, causal evaluation frameworks for LLM-driven agents, and observational-experimental data fusion at scale.
The AGI foundations research agenda: what’s real, what’s speculation, what I would invest in
An opinionated map of the contemporary AGI research landscape, separating genuine progress from benchmark inflation. Three categories (real and underfunded; real but saturated; speculative but worth hedging) with specific capacity allocations for a hypothetical AGI foundations lab.
Applied research and engineering
Production LLM agents, what the reliability engineering literature missed
Lessons from shipping production LLM agent systems, and a framework for agent reliability that extends beyond benchmark completion rates. Five concrete patterns from production: strict structured output, calibration-oriented prompting, constrained tool calls, deterministic retry chains, continuous shadow evaluation.
Quant finance in 2026, where ML actually helps, where it doesn’t
An honest assessment of where machine learning has improved quantitative finance (execution, credit scoring, exotic-option calibration, fraud detection) and where classical methods remain dominant (mid-frequency alpha, portfolio optimization, market risk). With specific time horizons and investment recommendations.
Quantum computing’s honest business case in 2026
Calibrated assessment of where quantum computing will deliver business value, when, and where the hype substantially exceeds the underlying technical reality. Recommendations for firms in cryptography, financial services, pharma/materials, and logistics.
Personal and leadership
From theoretical physics to applied AI: what transferred, what didn’t
An intellectual genealogy of the transition from theoretical high-energy physics to applied machine learning. What renormalization-group thinking taught me about inductive bias. What quantum field theory taught me about causal structure. Advice for physicists considering the transition and hiring managers evaluating ex-physicists.
What I would build in the first 90 days as a Research Director
A concrete 30/60/90-day plan for a hypothetical Research Director role: team onboarding, research-program selection, deliverable milestones, and external engagement. Read if you are evaluating my readiness for a research leadership role.
If any of these essays prompts an interest in talking about a specific role or collaboration, I am at iohanngrig@gmail.com.