RL is the new floor for senior execution researchers

The bar has moved

Execution research at top quant firms doesn't look like it did in 2023. The standard execution toolkit — TWAP, VWAP, implementation shortfall, basic impact models — used to define a senior execution researcher's competence. In 2026, that's table stakes. The differentiator is reinforcement learning, paired with deep TCA and microstructure intuition.

Since early 2026, every multi-strat platform and tier-one prop HFT firm we've worked with has either built or is building RL-based execution systems in production. Order routing, child-order placement, dynamic participation rates — all increasingly modelled as RL problems with reward functions tied to TCA outcomes rather than hand-tuned heuristics.

What this means for hiring

Candidates from a pure microstructure background — five years of optimization-heavy execution work, no ML — are no longer competitive for senior execution seats at the firms pushing the frontier. The bar has moved.

Conversely, ML researchers without microstructure intuition can't design the right reward functions or recognise when a strategy is exploiting transient liquidity.

The contested candidates are the rare profile that sits at all three: RL chops, TCA fluency, and microstructure intuition.