Third Street Systems
We work with life-science technology leaders who are rethinking how engineering gets done — from operating-model design to the engineering itself.
Kendall Square · Cambridge, Massachusetts
The shift
The last decade's operating model is giving way to something structurally different.
Large internal teams extended by outsourced capacity are giving way to small senior teams directing fleets of AI agents, with institutional knowledge captured rather than walked out the door.
That is not a tooling upgrade. It changes team shape, delivery economics, vendor strategy, and where senior judgment has to sit.
01
The model that got us here
A large internal team, extended by outsourced capacity when the roadmap outran it. For fifteen years this was simply how enterprise software got delivered — and it worked, because the constraint was hands on keyboards.
02
Where it breaks
Every added head adds edges, not just output. Context lives in people, so it leaves when they do — and in a validated environment, re-establishing it is not free. Most organisations are running new tools inside this shape and wondering why the leverage never arrives.
03
The new unit of productivity
Judgment concentrates. A handful of senior engineers who know the domain direct agent fleets that do the volume. The org chart gets shallower, the review bar gets higher, and the work that used to be bought by the head stops being bought at all.
04
What makes it hold
None of it survives contact with a regulated enterprise without a substrate underneath: captured context, evaluation you trust, and guardrails that hold. Get that layer right and institutional knowledge accrues instead of walking out the door.
What we do
We help leaders design that transition — and build through it.
Team shape, delivery economics, and vendor strategy for an engineering organisation where capacity is no longer bought by the head. Where senior judgment concentrates, what stops being outsourced, and how you sequence the change without stalling delivery.
The systems layer that makes the model real: how agents get context, what guardrails hold in a validated environment, how you evaluate output you did not write by hand, and how institutional knowledge compounds instead of leaving with headcount.
Building the thing. Working prototypes that prove the model on your problem and your data, and the production systems that carry it once the argument is settled.
Who you're working with
Founded by a former Novartis Executive Director who ran a 200+ person global engineering organisation through exactly this shift.
Software engineering, data engineering, devops, and enterprise architecture, across the Research division. We advise like operators, because we are: the same person who sits in the strategy review writes the code that proves it out.
We work at the executive table and in the codebase. Usually both in the same week.
Get in touch
If you are working through this transition — or deciding whether to start — a conversation costs nothing.
jeremy@thirdstreetsystems.com LinkedIn