Karan
Software Engineer specializing in Support Platforms and Community Tools within gUP.
Experienced in conducting deep-dive abuse and security audits to harden backend forum infrastructure against sophisticated evasion techniques. Actively pioneering zero-friction diagnostic flywheels and self-healing overlays using Google's frontier AI capabilities.
Constantly pushing for engineering excellence, system reliability, and secure, policy-compliant automation to scale Google's global community collective intelligence.
Research Areas
Authored Publications
Sort By
Preview abstract
Bidirectional (right-to-left, RTL) layout defects arise when interfaces authored with physical directional CSS (margin-left, border-left, text-align:left) are rendered under dir="rtl" without being mirrored. Accessibility engines (axe-core), actionability engines (Playwright), and even Stylelint's current core logical-property rule systematically fail to localise these defects. We present runtime structural invariants, a zero-golden mode over computed style and geometry, and a bidi-conservative differential mode — that catch them in a headless browser. On a held-out benchmark of 14 real RTL bugs reconstructed from dated fix commits in Bootstrap and Material Components Web, the detector recovers 13/14 (2 via zero-golden, 11 via the differential mode), with no false alarm on any patched version; on genuinely real historical compiled CSS it catches 1 of 2 attempted, using the actual committed bootstrap.rtl.css. On 16 negative LTR-island cases (8 hard) it reports 0 false positives, and both modes report 0 false positives on 43 unmodified real components across four design systems and three RTL architectures (rule-of-three 95% upper bound ≤ 7%). It surfaced a real directional defect : a physical text-align:left in .menu-list in the current Bulma 1.0.4. By contrast, Stylelint's core rule emits an identical 1744 warnings on Bootstrap's LTR and correct-RTL builds, pseudo-localisation detects none of the mirroring bugs, and a naive pixel mirror-and-diff reaches only F1 0.87 with a test-set-tuned threshold, whereas the differential needs none. We find the two modes complementary and scope this paper to detection; repair is future work.
View details
Preview abstract
Content moderation compresses a graded, contested human judgment into a binary act (leave up / take down) over billions of items, at a scale where running a frontier LLM on every post is uneconomical. A new class of frozen "System One" models returns typed, calibrated decisions instead of text at ~$0.000025 per call and ~100 ms. We run the first independent audit of one such model on 142,658 human-labeled comments, and benchmark it against 14 frontier reasoning models (the GPT-5.x, Claude, Gemini 3, and Grok families) on a shared probe set. Two findings stand out. First, the cheap model is accuracy-competitive: its zero-shot classification AUROC (0.83) is not statistically separable (paired bootstrap, n=56; a failure to reject parity, not proof of equivalence) from 13 of the 14 frontier systems while costing 1–3 orders of magnitude less, and it is no worse calibrated than several of them. Second, "confidence" conflates two uncertainties that come apart: human disagreement (where annotators split on satire and coded language) and decision risk (where the automatic action is costly-wrong). Under an asymmetric moderation loss these call for opposite escalation signals, so no single confidence number serves both. Because the model prices extra questions at nearly zero, decomposing the judgment into parallel sub-questions and reading their spread recovers a human-disagreement signal that the model's native confidence misses—but this remedy is specific: it helps the weakest, disagreement-blind models and only when the decomposition aligns with the sources of disagreement; a specialist toxicity model and the strongest LLMs do not need it, and the effect is distribution-sensitive. We release the instrument, the cached runs, and the analysis.
View details
