AI product builder with a 119-app portfolio across products, platforms, tools, and operating systems.

Run LLM spend like a professional operating discipline — FinOps for tokens.

129+ Claude SKILL.md prompts with copy-ready frontmatter.

Battle-tested prompts for task decomposition and autonomous reasoning.

Organize, version and share LLM prompts across teams.

A terminal UI for building, testing and refining prompts.

Visual workflow builder for agentic AI pipelines.

Generate automated business-process blueprints from plain language.

Design and orchestrate multi-agent AI systems.

Turn technical content into polished C-suite deliverables.

Social authority toolkit — content planning and thought-leadership frameworks.

CIT Alumni Association Chennai — AI articles hub.

Terminal-based creative AI tool for content and prompt engineering.

7-pillar AI Quality Engineering framework (ISO 42001, NIST AI RMF, EU AI Act).

AI-assisted test generation, execution and reporting framework.

Assess QE maturity, identify gaps and build improvement roadmaps.

Patterns, frameworks and reference architectures for modern enterprises.

Frameworks and maturity models for scaling Global Capability Centers.

Operating dashboard for GCC leaders — maturity, lifecycle and AI advisor.

Technology evolution, migration paths and modernization strategies.

Methodology for discovering, filtering and sharing high-signal tech content.

Go-to guide for healthcare GCC players, professionals and leaders.

Pharma digital transformation and GCC strategy guidebook.

A structured clinical reference making oncology knowledge searchable.

1,120+ prompts with a Bloomberg-style FINPROMPT terminal.

Personal expense tracking with budgets, trends and export.

Courses, quizzes and learning paths for enterprise professionals.

AI learning and terminology hub.

Design and track enterprise learning and competency programs.

Course materials, facilitator guides and program designs.

Interactive lessons and simulators for quantum computing concepts.

Quiz platform across enterprise domains and emerging tech.

Fast-paced timed quizzes with leaderboards.

Cybersecurity lexicon, frameworks and threat taxonomies.

Curated competitions, learning paths and data-science walkthroughs.

Structured learning paths across domains.

Interactive learning app for 10th-grade students.

One-stop resources and prompting strategies for scholars.

A studio for drafting and refining academic research papers.

Searchable record: 482 certifications, publications, awards.

Tamil Nadu election prediction tool.

Swiss-army utility hub of essential developer and productivity tools.

Plain-language explanations and analogies for complex tech.

AI matcher: analyze job descriptions, generate tailored profiles.

Professional profile search and comparison across domains.

Curated questions, frameworks and practice scenarios.

App ideas with feasibility analysis and build roadmaps.

Centralized launchpad connecting all KR micro-apps.

News aggregation with category filters and AI summarization.

A pocket news reader.

A unified social dashboard.

Retro browser-based arcade games collection.

Classic Tetris with modern UI and scoring.

Jeopardy-style quiz game.

A calm, mindful focus companion.

Duas with Arabic, transliteration, translation and context.

150 technique-correct prompts that turn any LLM into a quality-engineering co-pilot — built around real test-design technique, not generic prompt tips.

120 SDET interview questions answered with the reasoning, plus a certification navigator and role roadmaps.

Paste a user story and acceptance criteria, get a structured test suite. Deterministic engine plus bring-your-own-key AI — your key stays in your browser.

Upload JUnit XML and get your flaky tests ranked and explained — not just detected. Results never leave your machine.

The prompt manager for people who've outgrown a notes app: library, variables, versions, one-click run in Claude or ChatGPT.

Realistic synthetic test data from any schema — PII-safe by construction, so nothing generated can match a real person.

A 5-year fully-loaded TCO model for capability-centre location decisions — including the attrition replacement cost vendor decks leave out.

Score your QE maturity across six weighted dimensions in five minutes, then get the ROI business case a CFO will actually read.

Requirements in, first-draft IQ/OQ/PQ protocols out — with GAMP 5 categorisation and a Part 11 gap checklist. A draft accelerator, never a system of record.

A sequenced certification roadmap across 39 certs and 8 roles — with the renewal burden nobody plans for, and honest verdicts on which ones to skip.

Nine rungs from first Kaggle submission to medals, each with the failure mode that stalls most people there. From a Legacy Grandmaster.

Paste an HL7 v2 message and see exactly how it maps to FHIR R4 — a real parser that refuses to guess when a value has no honest equivalent.

Assess your AI governance readiness, then draft the artifacts a regulated organisation needs — acceptable-use policy, validation SOP, oversight protocol, board memo. Templates to adapt, not legal advice.

Self-audit against 43 controls from 21 CFR Part 11, EU GMP Annex 11 and ALCOA+ — scored on the evidence you actually hold, with gaps ranked by inspection exposure.

Answer seven questions about your project and get a finished test strategy and test plan — coherent, Word-ready, with the decisions on page one instead of page nineteen.

Paste a CARC denial code and get the plain-English meaning, the root causes worth checking first, who owns the fix, and whether it is worth appealing at all.

Twenty observation themes drawn from FDA Form 483s, each with the evidence that answers it, the questions an inspector uses to get there, and a readiness score across seven areas.

Answer six questions and get a capability-centre operating model: archetype, org shape, an HQ-to-centre RACI with one accountable party per domain, governance cadence and KPIs for your stage.

Pick your current role and your target role across four tracks and get the real skill gap, the next three moves, and the reasoning behind any cross-track jump — free. No computed paths, only real, mapped moves.

Answer six questions — language and tool, CI, reporter, BDD, parallelism, Docker — and download a clean, runnable Playwright + TypeScript automation framework, zipped in your browser. Every file is authored, not generated.

Add the certifications you hold and see exactly how many CE credits you still owe, the pace per month to finish on time, and days to every deadline — with exam-recert and lifetime credentials handled honestly. The rules are editable defaults; the math is always exact.

Paste your CI and test numbers as CSV and get an exec-ready board: the four DORA keys beside four quality signals — 8 metrics in all, each banded, trended with a direction-aware sparkline, and folded into an auto-written executive summary. Everything runs in your browser.

Graded practice at writing prompts for real QE tasks — boundary tests, a bug report from a vague ticket, API negative cases, flaky-test triage. Write the prompt, grade it against an authored checklist of what a strong prompt carries, and compare with the model answer. It detects signals, never judges meaning.

Is your software a medical device — and what IMDRF risk category? A plain scope screen, then category I–IV read off the public IMDRF N12 matrix from the state of the healthcare situation and the significance of the information. An educational aid over a real framework, not a regulatory determination.

Fill in what you know about a decision and get an answer-first board/exec deck outline — BLUF, situation, options, recommendation, risks, the ask — graded against a 12-point completeness rubric that names what a board will stall on. It structures your inputs; it does not write your deck or invent facts.

Answer what you've decided about AI at your organisation and get a draft AI acceptable-use policy assembled from an authored 22-clause library — with the core topics you haven't decided flagged as gaps and contradictions in your answers flagged as conflicts. A template starting point, not legal advice.

Which LLM for the job is a weighting problem. Set how much reasoning, cost, latency, context, coding, tool use and multimodal matter, add hard constraints like open-weights, and get a ranked shortlist of model classes with the reason for each — from an authored, dated matrix. Not live pricing; verify with the provider.

Describe the Global Capability Centre you want to stand up and get a plan that sequences the real workstreams — strategy, entity, real estate, talent, technology, operating model, transition — against their dependencies, on a timeline scaled to your target, with the critical path and your undecided calls flagged. A planning aid, not a delivery guarantee.

A personal knowledge system that runs entirely in your browser: capture notes, connect them with [[wiki-links]] and automatic backlinks, tag and search them, and let old notes resurface for review. Your notes stay on your device — nothing is uploaded — and export any time.

Bring-your-own-syllabus spaced-repetition flashcards that run entirely in your browser. Make cards from anything you're learning (or paste a whole syllabus as 'front | back' lines) and review your due cards on an SM-2 schedule that shows each just before you'd forget it. Your cards stay on your device — nothing is uploaded — and export any time.

Pick a role, level and city and get a directional pay band from an authored, dated matrix — a base band scaled by transparent level and city multipliers, all in USD. Compare two cities on nominal AND cost-of-living-adjusted pay, so you see which offer actually buys more. Directional, not a live feed — verify before you rely on it. 100% in your browser.

A behavioural-interview story bank with competency-coverage analysis. Write your own STAR stories, tag them to competencies, and see which of the common question archetypes you can answer strongly and which are gaps. It authors nothing and predicts no specific question — it just makes sure you walk in prepared for everything they tend to ask. 100% in your browser.

Running an AI capability in a GCC is a rhythm, not a project. Assemble your operating instruments — the recurring reviews and controls an AI function needs — from an authored catalogue of 22 instruments across 7 domains, set each one's cadence and owner, and see your operating calendar, the dark domains with no active control, and the instruments running with no owner. It produces no score — a coverage checklist, not a maturity verdict.

A number only means something next to a benchmark. Pick your region, enter your capability centre's own org-level figures — cost per FTE, span of control, seniority mix, attrition and more — and each is placed against an authored, dated band so you see where you're comfortably in range and where you're an outlier. 7 metrics across 4 regions. Directional 2026 ranges, blended across roles — a reference to pressure-test, not a live survey, and not a pass/fail bar.

A regulatory review register, not a live feed. Keep the regulations that matter to you in one place — mark what applies, who owns it and when you last reviewed it — and see which reviews are overdue, which regs you've never assessed, and which have nobody accountable. 16 regulations across 6 domains, an authored dated catalogue (2026-08); verify currency yourself. 100% in your browser.

Paste two versions of an OpenAPI 3.x spec and get a deterministic diff — every change categorised breaking, risky or safe, with the reason for each. The same heuristics real contract-diff tooling uses. Reads JSON, not YAML; doesn't resolve $ref. 100% in your browser.

Answer twelve yes/no questions about your CI/CD pipeline and get it scored against an authored anti-pattern checklist — the same items that keep showing up in CI/CD post-mortems — with a worst-first attention list and a sequenced fix order. Not a config parser. 100% in your browser.

Paste your test titles and get near-duplicate clusters flagged by token-overlap similarity — a real, deterministic signal for a suite audit, worst pairs first. Not semantic AI: a similarity signal to point a reviewer at, not a verdict. 100% in your browser.

Answer five questions about a task and get a recommended agentic architecture — from an authored decision tree — with the failure modes and guardrails that pattern needs, and why the other five don't fit as well. No live agent, no templates to import. 100% in your browser.

Paste your agent's system prompt and see which of eight authored guardrail categories it addresses and which it's silent on. A checklist-based gap scan, not a penetration test or a guarantee. 100% in your browser.

Pick your task type and the failure modes you're worried about, and get a starter eval-question scaffold plus a rubric structure from authored patterns — you fill in your feature's expected answers. Runs no evaluation itself. 100% in your browser.

Set weighted criteria and score named vendor proposals against them — get a ranked comparison, each vendor's weakest link, and an 8-item red-flag checklist that stays visible next to the rank, never averaged into it. Downstream of a sourcing decision, not a replacement for one. 100% in your browser.

Pick which of 16 authored services your capability centre offers across 6 categories, set committed response hours, and see exactly which ones have no SLA at all — gaps sort first, for visibility. An authored, dated reference band to pressure-test against, not a benchmark claim. 100% in your browser.

Enter a business case — upfront cost, recurring cost, hard savings, soft benefit, risk avoidance, periods, discount rate — and get discounted cash flow NPV, simple payback and a sensitivity table, plus the 8-question rebuttal checklist a board reliably asks, free. 100% in your browser.

Log Corrective and Preventive Actions and get a register showing overdue effectiveness checks and aging CAPAs first, plus a 7-item maturity score computed against your own data, drawn from ICH Q10 / ISO 13485 CAPA principles. What happens after a finding, which nothing else here covers. 100% in your browser.

Pick a payer type and one or more procedure categories and see whether prior authorization is typically required, common documentation needed, and typical turnaround — before you submit, from an authored, dated baseline. No CPT codes, no PHI. 100% in your browser.

Paste a job description and your resume and see, from an authored 97-term dictionary, which skills/tools/certifications the JD mentions and which your resume is missing. Deterministic string matching, not an AI rewrite tool. 100% in your browser.

Enter your savings, severance, monthly expenses, COBRA cost and unemployment eligibility and get a month-by-month runway simulation, plus a 16-week job-search milestone framework compared against your actual runway. 100% in your browser.

Log courses, certs, books and conferences with time invested and a self-rated impact, and get a portfolio view of time-by-category, a high-hours-low-impact flag, and a gap check against the skills you're actually targeting. 100% in your browser.

Log professional contacts with a relationship-strength rating and last-contact date, and get a recency-weighted who-to-reach-out-to-this-week list plus templated reconnection and warm-intro messages filled from your own details. 100% in your browser.

Describe your agent's call pattern — calls per task, tokens per call, model tier, any tool costs — and get cost-per-task and cost-at-volume figures, plus a 6-item cost-reduction lever checklist. Downstream of a model choice, not a replacement for ModelMatch. 100% in your browser.

An authored, categorised library of 35 adversarial test prompts across 7 categories to manually red-team your own agent, with a self-recorded pass/fail session tracker. Defensive testing, no live execution. 100% in your browser.

Pick a denial argument category, enter the supporting facts, and get a structured appeal letter assembled from an authored clause library — with missing recommended fields flagged, not silently dropped. The document DenialDecoder deliberately doesn't draft. 100% in your browser.

Enter headcount, attrition rate, average salary and a replacement-cost multiplier, and get the total annual cost of attrition in dollars — plus a side-by-side ROI comparison against a proposed retention investment. 100% in your browser.

Paste a tool/function-calling JSON schema and get it checked against 7 authored structural checks: named and identifier-safe, a real description, typed and described parameters, and a required list that actually matches your properties. 100% in your browser.

Enter the payer's allowed amount, remaining deductible, coinsurance percentage or a flat copay, and the remaining out-of-pocket maximum, and get a full patient-responsibility breakdown with the total correctly capped at the OOP max. 100% in your browser.

Enter your model's context window, system prompt size, tool schema size, reserved output and average turn size, and see the fixed overhead, usable history budget, and how many turns fit before truncation risk. No authored context-window table. 100% in your browser.

A deviation just landed and nothing is open yet. Answer 8 triage factors and get a minor/major/critical classification, the investigation depth it warrants, the 6 hard floors that override a raw score, and the inputs that contradict each other. A disclosed rubric, never a regulatory classification. 100% in your browser.

Every agent capability sits somewhere between "just let it run" and "two people must sign". Describe one capability across 9 factors and get the approval gate it warrants on a 6-rung ladder, the 7 hard floors that override a raw score, the controls it implies, and the answers that contradict each other. A disclosed rubric, never a security audit. 100% in your browser.

A claim that isn't denied can still be paid wrong. Paste your EOB lines against your own contracted rate and see which were allowed below contract, classified into 5 triage families — a rate mismatch worth disputing, bundling, a claim correction, or a real denial that was never a rate question. 100% in your browser.

An agent already did something wrong — or something that could have. Answer 9 triage factors and a root cause, and get a minor/major/critical classification, the investigation depth it warrants, the 8 hard floors that override a raw score, and a root-cause-specific focus checklist. 100% in your browser.

A planned change needs sizing before it is made, and the expensive mistake runs one way: a change called “like-for-like” that turns out to have moved a validated process. Answer 8 factors and get a minor/moderate/major impact tier, the approvals it needs, the revalidation it triggers, and the 8 hard floors that override a raw score. 100% in your browser.

Preventing a denial is a different job from working one. Paste your claim queue and see which lines have a known, fixable problem on them right now — including 2 hard blocks that will be denied on receipt — across 7 disclosed checks. No CPT or ICD content. 100% in your browser.

A binary gate is why release meetings go badly — the room negotiates its way to “go” and the concerns become nothing. Enter the evidence for one release and get go, go-with-conditions or no-go, the 8 hard floors that override a raw score, and 8 conditions each with a deadline attached. 100% in your browser.