Engineering
Ready to connect with your peers in engineering? Grab your free ticket for Startup Boston Week on Sept 14 - 18 today!
Vibe Coding's Hangover: When the Tech Debt Bill Comes Due
AI-generated code has dramatically lowered the barrier to building software. Founders are launching products faster, testing ideas sooner, and shipping features that would have taken months to build just a few years ago.
But speed comes with tradeoffs. As products gain users, onboard engineers, undergo security reviews, and evolve into real businesses, many teams discover that code written quickly can become increasingly difficult to maintain, secure, and scale.
This session explores where AI-generated code creates leverage, where it introduces risk, and how teams can balance speed with long-term sustainability.
You will learn:
-
The most common failure points in AI-generated and vibe-coded applications
-
Warning signs that technical debt is becoming a business risk
-
Practical ways to introduce engineering rigor without slowing innovation
-
Security considerations and audit strategies for AI-generated codebases
-
How to decide whether to refactor, rebuild, or continue investing in an existing codebase
The goal isn't to stop moving fast. It's knowing which shortcuts help you reach product-market fit and which ones create problems you'll eventually have to pay for.
Join us for an honest discussion about the tradeoffs behind AI-assisted software development.
Trust Isn't a Feature: Engineering Products Users Actually Trust
The best products don't just solve problems. They create confidence.
Whether you're building software, hardware, or AI-powered experiences, user trust often determines adoption, retention, and advocacy more than any individual feature.
The challenge is that trust isn't built through marketing claims, it's earned through product decisions, user experience design, transparency, reliability, and giving users a sense of control when things don't go as expected.
This session explores how product teams intentionally design for trust, from privacy-first experiences and feedback loops to explainability, confidence signals, and human oversight.
You will learn:
-
The UX patterns that consistently increase user confidence and trust
-
How to design privacy-first experiences without creating unnecessary friction
-
Approaches for building transparency and control into AI-powered products
-
Strategies for identifying and mitigating bias, edge cases, and trust-breaking experiences
-
Methods for measuring trust and understanding whether product changes are strengthening or weakening it
Trust isn't built through a single feature or policy. It's the result of hundreds of product decisions that help users feel informed, empowered, and confident. Join us to learn how leading product teams design trust into their products from day one.
From Context to Control: Scaling Agentic AI Systems Beyond the Demo
Most teams can get an AI agent to work in a demo. The real challenge begins when that agent has to operate reliably in production.
As agentic AI systems move from experimentation to deployment, teams encounter a new set of problems: context windows break down, agents get stuck in loops, latency becomes unpredictable, costs spiral, and debugging feels nearly impossible.
This session explores the engineering and product decisions that separate impressive demos from production-ready systems that can scale.
You will learn:
-
Common failure modes in agentic AI systems and how to identify them early
-
Architectural patterns for building reliable, observable, and debuggable agent workflows
-
Strategies for managing context, memory, latency, and cost as systems scale
-
Approaches for monitoring agent behavior and diagnosing failures in production
-
How successful teams move from reactive troubleshooting to intentional system design
Building an agent is easy. Operating one at scale is hard. Join us to learn how experienced teams design agentic systems that remain reliable, performant, and cost-effective long after the demo ends.
Physical AI: Engineering Intelligence That Moves
For the past few years, most AI innovation has lived on screens. Now, foundation models are beginning to move into the physical world.
From humanoid robots and autonomous systems to intelligent manufacturing, logistics, healthcare, and industrial applications, physical AI is creating new opportunities and entirely new engineering challenges.
This session explores what happens when AI leaves the browser and interacts with the real world, covering the technologies, infrastructure, and breakthroughs that are making physical AI possible today.
You will learn:
-
What recent advances have accelerated the development of physical AI systems
-
The engineering challenges that make robotics and physical AI fundamentally different from software products
-
What it takes to build and fund a physical AI startup
-
Why Boston has emerged as a global hub for robotics and physical AI innovation
-
The technical and commercial bottlenecks still limiting adoption and what comes next
Building software is hard. Building intelligence that can safely interact with the physical world is even harder. Join us for a conversation about the future of robotics, autonomy, and the next generation of AI-powered systems.
Built to Ship: Designing Hardware Products That Survive the Factory Floor
Most hardware products are designed to impress in a demo. Far fewer are designed to survive manufacturing.
The decisions made during product design have ripple effects across sourcing, assembly, quality control, cost, and scalability. A feature that feels elegant in a prototype can create headaches on the factory floor, while a seemingly small design choice can determine whether a product launches on time, on budget, and at scale.
This session explores how designers and hardware teams can incorporate manufacturing realities earlier in the process without sacrificing user experience or product vision.
You will learn:
-
How to design with manufacturing constraints in mind from the start
-
Ways to balance usability, aesthetics, cost, and manufacturability
-
Common design decisions that create expensive downstream problems
-
Frameworks for making smarter product tradeoffs before entering production
-
How successful hardware teams collaborate across design, engineering, and manufacturing
The best hardware products aren't just designed for users. They're designed to be built, tested, assembled, and shipped repeatedly at scale. Join us to learn how great teams bridge the gap between product vision and manufacturing reality.
