Skills

Designs a customer churn prediction approach end to end: framing the target and prediction window, engineering behavioral and tenure features, choosing suitable models and evaluation metrics for imbalanced data, and planning how predictions feed retention actions. Focuses on leakage-free design and metrics that match the business goal.

Analyzes Terraform plan output and state to find configuration drift between declared infrastructure and what actually exists. Explains each drifted resource in plain language, flags risky or destructive changes, and recommends whether to reconcile in code, import, or apply, so teams can keep state and reality aligned safely.

Writes clear, calm customer-facing incident communications during an active outage: status-page updates, in-app banners, and stakeholder emails. Adapts the message to the incident phase (investigating, identified, monitoring, resolved), states impact and scope honestly without leaking sensitive detail, and keeps a consistent tone across updates.

Diagnoses failing Kubernetes workloads from symptoms, events, logs, and manifests. Works through a systematic checklist (pod status, events, resource limits, probes, scheduling, networking) to identify the root cause of issues like CrashLoopBackOff, ImagePullBackOff, OOMKilled, and pending pods, then recommends concrete fixes and verification steps.

Turns notes, articles, or textbook passages into spaced-repetition flashcards. Produces atomic question-and-answer pairs (and optional cloze deletions) that test one concept each, avoiding ambiguous prompts, and can export in a plain format ready to import into tools like Anki.

Drafts accurate, well-supported answers to vendor security and compliance questionnaires (SIG, CAIQ, custom due-diligence forms). Maps each question to existing policies and control evidence, writes clear responses, and flags any question where the honest answer is unknown or a control is missing rather than overstating posture.

Guides a git bisect session to isolate the commit that introduced a regression. Helps define a reliable good/bad test, drives the bisect steps, interprets results, and can suggest an automated test script for `git bisect run`, then explains the offending change once the culprit commit is found.

Generates JSON Schema (Draft 7 through 2020-12) from example payloads, TypeScript types, or a natural-language description of a data shape. Infers types, required fields, formats, and constraints, and can add descriptions and enums to produce a validation-ready schema.

Drafts tailored, concise cover letters that connect a candidate's experience to a specific job description. Extracts the role's key requirements, maps them to the applicant's most relevant achievements, and produces a focused letter in the requested tone without generic filler.

Synthesizes user research, interviews, and survey data into clear, evidence-based user personas. Clusters behaviors and needs, names each persona, and captures goals, pain points, motivations, and representative scenarios, keeping the output grounded in the source data rather than invented detail.

Finds, validates, and fixes security vulnerabilities end to end, going beyond audit-only review to produce verified patches. Reproduces the issue, proposes a minimal fix, and confirms the vulnerability is closed without breaking existing behavior. Models a CodeMender-style workflow where multiple agents detect, validate, and patch code security issues at scale.

Designs and drives computer-use agents that operate a desktop or browser directly via screenshots and UI actions (click, type, scroll, navigate). Plans multi-step GUI workflows, handles verification and recovery between steps, and structures guardrails for safe autonomous execution. Built for models exposing computer use as a native client-side tool.

Designs durable memory for AI agents, including what to retain, how to retrieve it, privacy boundaries, expiration policies, and evaluation criteria for useful recall.

Produces blameless, actionable incident postmortems with a precise timeline, contributing factors, impact, decisions, corrective actions, owners, and measurable follow-through.

Reviews Model Context Protocol servers for excessive permissions, unsafe tool design, prompt injection exposure, secret handling, authorization gaps, and auditability.

Synthesizes interviews, support tickets, reviews, and survey responses into evidence-based themes, opportunities, representative quotes, confidence levels, and product actions.

Designs traces, metrics, structured events, privacy controls, dashboards, and alerts for LLM applications so teams can diagnose quality, cost, latency, and tool-use behavior.

Tests AI agent tool use with realistic fixtures, malformed inputs, permission boundaries, deterministic assertions, and recovery checks for failed or partially completed actions.

Gerçekçi uçtan uca senaryolarda görev başarısını, araç kullanım doğruluğunu, dayanaklılığı, gecikmeyi, maliyeti ve güvenli toparlanmayı ölçen yeniden üretilebilir AI agent değerlendirmeleri oluşturur.

Diagnoses failing browser and end-to-end tests by correlating assertions, traces, screenshots, network activity, timing, and application state into reproducible fixes.

Optimizes prompts and agent workflows for finite context windows by prioritizing evidence, compressing history, managing retrieval budgets, and measuring token-cost trade-offs.