Tech News
← Home  ·  All topics

Engineering

224 GoKawiil briefs on this topic · RSS for engineering

Detection-as-Code Foundry pairs LLM agents with a deterministic referee to validate Sigma rules

A new multi-agent system called Detection-as-Code Foundry uses four role-specialized LLM agents—Threat Analyst, Blue Detection Engineer, Telemetry, and Red-Team Adversary—built on LangGraph, Pydantic, FastMCP and Claude to turn a threat input like a CVE or ATT&CK technique into a Sigma detection rule. The rule is run through an adversarial Blue-vs-Red loop and shipped with a proof bundle: a positive and benign test corpus, an adversarial mutation set, and a coverage label (real-coverage, brittle, or broken) with FP/FN counts and a reproducibility hash. A deterministic engine, not the LLMs, decides all pass/fail verdicts, evasion results and the deployment gate.

NFL QB Josh Dobbs builds aviation venture inspired by NASA work

Detroit Lions backup quarterback Josh Dobbs, an aerospace engineering graduate of the University of Tennessee, says he is developing an aviation infrastructure business aimed at improving how athletes travel. Dobbs, nicknamed 'The Passtronaut,' credits his engineering training and past NASA-related work for shaping the idea, drawing parallels between problem-solving in football and engineering.

IEEE Computer Society Examines Accessibility Challenges in Pervasive Computing

IEEE Computer Society published commentary describing how computing tasks now span phones, voice assistants, wearables and connected devices, often triggered by location or movement rather than direct app use. The piece frames this shift as raising new accessibility engineering questions for designers building across multiple interfaces and contexts.

Jane Street adds new interview rounds testing design skills and AI-assisted coding

Jane Street says it is trialing new software engineering interview formats for campus candidates, starting this summer. These rounds involve less hands-on code writing and more focus on software design, code quality verification, and using LLMs to produce code, while most rounds remain traditional coding interviews done without AI assistance.

Engineering team drops peer code review, lets developers approve their own AI-written PRs

A software team eliminated traditional peer code review after struggling to keep pace with a growing volume of pull requests. Instead, developers now work with Claude to generate code and open PRs, then review and merge their own AI-assisted work once CI checks pass, rather than waiting for a second engineer's sign-off.

UC Irvine seminar presents research on gap between AI coding tool use and software output

A UC Irvine software engineering talk on October 30, 2026 reviewed several studies examining why increased use of AI coding assistants has not translated into faster software shipping. The talk cited research including an 'Activity-Outcome Gap' paper, an NBER working paper comparing productivity across AI coding tool generations, and studies on code review agents and developer trust in AI autonomy. The 21-minute recorded talk is available with captions and as a slide-by-slide version online.

Researchers apply complex-network principles to engineer mechanical metamaterials

Writing in Nature, Both et al. describe a method for designing materials whose internal structure mimics the connectivity patterns found in complex networks, such as variable node connections and component sizes. The approach borrows concepts established by network scientists Albert-László Barabási and Réka Albert to produce materials with notable mechanical properties.

New guide frames inference engineering as the next platform-team skillset

A field guide aimed at platform engineers outlines why organizations are shifting to running open-source AI models themselves rather than relying solely on hosted frontier models. It cites DeepSeek's R1, released in January 2025, matching OpenAI's o1 on math and coding benchmarks within about a month, and notes Hugging Face now hosts over two million open models. The guide also references Upbound's Project Champagne work running models on its Modelplane tool, where most engineering effort went into infrastructure like TLS, identity and observability rather than model serving itself.

AI tools used to reverse-engineer commercial software and Turla malware

Researchers demonstrated that AI-assisted methods can unpack, map and recreate the internal workings of complex software, including the Turla malware family. The work shows reverse engineering tasks that once required specialized expertise and significant time can now be accelerated substantially using AI tools.

GitHub Next engineer Maggie defines the 'design engineer' role at Granola meetup

Speaking at a Granola-hosted meetup for design engineers, Maggie, a Staff Research Engineer at GitHub Next, described her career path through illustration, art direction, UX/UI and product design before landing in a hybrid research-engineer role. She argued that design engineering should not be reduced to flashy micro-interactions, but instead means designers who understand software's technical materials—state, data flow, databases, APIs, auth and latency—and who act as a bridge between design and engineering teams.

Australia's National Industry PhD Programme produces first graduate

Yifan Wang, a PhD student in electrical engineering and telecommunications at the University of New South Wales, submitted the first thesis completed under Australia's National Industry PhD Programme in May. The programme pairs students with university and industry co-designed research projects, backed by government and industry funding, and has so far funded 448 awards after more than 1,000 applications across seven rounds.

Essay Argues Engineers Still Outperform Peers Despite AI Agent Takeover of Coding

A software engineer writes that AI agents now handle most coding, debugging, documentation, and even architecture work, leaving engineers mainly coordinating multiple agents so they don't block each other. Using a proof-by-contradiction argument, the author notes that if humans added no value, skilled and average engineers would produce similar results with the same models and tools — but the gap between them persists, and top performers have become even more productive.