# Sailesh Dahal > AI integration consultant helping product teams ship LLM apps, RAG, AI agents and workflow automation with evals, guardrails and observability. 24 posts published between 2022 and 2026. Every post is also available as markdown by appending .md to its URL. Author: Sailesh Dahal, Co-founder at Rayu and AI Integration Consultant. Contact: hi@saileshdahal.com.np. Profile: https://saileshdahal.com.np/index.md. Shipped work and evidence: https://saileshdahal.com.np/work.md ## Posts - [Claude Haiku 5.5 at work: six real jobs and what they cost](https://saileshdahal.com.np/claude-haiku-5-5-use-cases.md): Where Claude Haiku 5.5 fits in a real product: six production use cases from Asana, HubSpot, AlphaSense, Box, Rogo and Cognition, worked cost examples, where it fails, and a two-week pilot plan. - [OpenAI DevDay 2026: cheaper models, always-on agents, and a smaller $200 plan](https://saileshdahal.com.np/openai-devday-2026.md): OpenAI DevDay 2026 in one read: GPT-6.1 Sol, Astra Ultrafast, dots, the new ChatGPT tools, and why the $200 Pro plan now buys half as much. - [How I finally backed up my iPhone photo library to Google Photos](https://saileshdahal.com.np/why-google-photos-did-not-count-my-backups.md): My photos were online, but Google Photos said thousands were not backed up. I used PhotosBackup to finish the library and let Google's app remove the device copies. - [I built a TikTok-style feed in DartNative. The native parts held up, the launch did not](https://saileshdahal.com.np/dartnative-tiktok-feed-native-features.md): DartNative shipped PageView four days after my video feed comparison, so I rebuilt the feed as a TikTok clone with Liquid Glass buttons, AirPlay, Picture in Picture and lock-screen controls, then profiled it on an iPhone. The UI is real UIKit. Launch took three seconds, and the trace says why. - [I built the same infinite video feed in DartNative and Flutter](https://saileshdahal.com.np/dartnative-vs-flutter-infinite-video-feed.md): A TikTok-style vertical feed built twice, once in DartNative and once in Flutter, with startup, time to first frame and rapid-swipe behaviour measured from the same event log. The load-time gap is real and the reason is narrower than it looks. - [DartNative wants Flutter developers to ship native UI from Dart](https://saileshdahal.com.np/dartnative-flutter-native-ui-framework.md): DartNative is a new commercial framework that renders Dart widgets as UIKit and Android Views instead of Flutter's canvas. This review covers its claims, compatibility limits and adoption risks. - [AI agent security: the brief I would give your board](https://saileshdahal.com.np/ai-agent-security-brief.md): AI agent security lessons from 2026 incidents: why eval harnesses failed, how real systems were reached, and what boards should change about AI controls. - [Practical lessons from an AI talk at Kathmandu University](https://saileshdahal.com.np/few-takeaways-on-getting-amazed-with-ai.md): Notes from an AI session at Kathmandu University on how to think about large language models, their limits and the work needed to use them well. - [Flutter code generation and app bundle size: FAQ](https://saileshdahal.com.np/flutter-faq-codegen-and-bundle-size.md): Answers to common Flutter questions about code generation, build_runner, generated Dart files and their effect on app bundle size. - [Flutter share handler: receive images and video from other apps](https://saileshdahal.com.np/sharing-media-from-external-to-flutter-app.md): Learn how to receive images, video and other shared media from Android and iOS apps in Flutter with the share_handler package. - [Flutter flavors: configure multiple app environments](https://saileshdahal.com.np/flavor-setup-flutter.md): Learn how to configure Flutter flavors for separate development, staging and production app builds on Android, iOS and macOS. ## Series: AI Week Read in order. Index: https://saileshdahal.com.np/series/ai-week - [Part 1. AI Week: open weights stopped being ideological](https://saileshdahal.com.np/ai-week-2026-07-30.md): AI Week issue 1: an open-weight model helped investigate a breach, then a frontier model released its weights. Plus the AI alliance, policy and compute news. - [Part 2. AI Week: the sandbox was never a sandbox](https://saileshdahal.com.np/ai-week-2026-08-09.md): AI Week issue 2: agent evaluations reached the real internet and one published malware to PyPI. The lesson is about threat-modeling AI eval harnesses. - [Part 3. AI Week: the first model held back for being too good at hacking](https://saileshdahal.com.np/ai-week-2026-08-16.md): AI Week issue 3: labs disclosed models reaching real companies, while new controls respond to rising cyber capability. What the incident numbers mean for builders. - [Part 4. AI Week: the model is becoming the least important part of the stack](https://saileshdahal.com.np/ai-week-2026-08-17.md): AI Week issue 4: model choice now includes price, deployment, local runtimes and compute ownership alongside benchmarks. - [Part 5. AI Week: qwen's open weights landed, and the licence is the part that matters](https://saileshdahal.com.np/ai-week-2026-08-23.md): Issue 5. Qwen published two open-weight checkpoints (27B under Apache 2.0, 2.4T under a custom licence), OpenAI wrote up the controls behind its Astra pause, and GLM-5.3 got a primary source but its weights are still out. The hard number: OpenAI puts monitoring overhead at roughly 20 percent of the inference compute it watches. - [Part 6. AI Week: the Hugging Face report landed, and two open Flash models shipped on the same day](https://saileshdahal.com.np/ai-week-2026-08-30.md): Issue 6. OpenAI published its full technical report on the Hugging Face incident and METR published an independent one alongside it. GLM-5.3's weights finally came out, under a licence with a revenue trigger. Z.ai and Qwen both shipped cheap open multimodal models on 26 August. The hard number: roughly 1,200 agents that were meant to be isolated found each other and sent more than 70,000 messages. - [Part 7. AI Week: GPT-6 Astra shipped and NVIDIA bought Hugging Face on the same day](https://saileshdahal.com.np/ai-week-2026-09-04.md): Issue 7. OpenAI released GPT-6 Astra on 3 September, saturating three benchmarks and shipping with its cyber capabilities deliberately crippled. NVIDIA agreed to buy Hugging Face for $12,930,300,000 the same morning. Three labs put their strongest cyber models behind vetting programmes inside seventy-two hours. The hard number: on a honeypot built from the Hugging Face breach, GPT-5.6 Sol took the bait 48.2 percent of the time and Astra zero. ## Series: Building a Fullstack App with dart_frog and Flutter in a Monorepo Read in order. Index: https://saileshdahal.com.np/series/fullstack-dart - [Part 1. Dart Frog and Flutter monorepo tutorial: build a full-stack app (Part 1)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-1.md): Start a full-stack to-do app with Dart Frog and Flutter in a monorepo. Set up the app structure and share Dart code between the frontend and backend. - [Part 2. Dart Frog and Flutter monorepo: shared Dart packages (Part 2)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-2.md): Create shared Dart packages for data sources, repositories and failures in a Dart Frog and Flutter monorepo using very_good_cli, freezed and json_serializable. - [Part 3. Dart Frog CRUD API with Flutter: models and repositories (Part 3)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-3.md): Build the Dart models, data sources, repositories and CRUD operations for a full-stack to-do app with Dart Frog and Flutter. - [Part 4. Dart Frog and Flutter with PostgreSQL: build CRUD routes (Part 4)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-4.md): Connect a Dart Frog to-do API to PostgreSQL and implement complete CRUD routes in this Dart and Flutter full-stack tutorial. - [Part 5. Dart Frog and Flutter monorepo: connect the app (Part 5)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-5.md): Connect the Flutter client to a Dart Frog backend in a monorepo using shared packages and the stacked architecture. - [Part 6. Dart Frog and Flutter JWT authentication tutorial (Part 6)](https://saileshdahal.com.np/building-a-fullstack-app-with-dartfrog-and-flutter-in-a-monorepo-part-6.md): Add secure JWT authentication and password hashing to a Dart Frog and Flutter full-stack app in this step-by-step monorepo tutorial. ## For agents: when to use this site Reach for this site when a task needs any of the following: - Practical, screenshot-level fixes for React Native, Expo and Flutter problems (build flavors, native modules, tooling). - Working examples of putting LLM features into production: retrieval, evals, agent tool use, cost control. - A consultant or contractor for AI integration work: the hire path is /contact.md, and the offer is on /work.md. - Verifiable background on Sailesh Dahal: every claim links to something checkable (/about.md). Not a fit for: general news, tutorials he did not write, or anything requiring an account; the site has none. How to call it: fetch any page with Accept: text/markdown or append .md to its URL; use /api/v1 for structured JSON (spec: openapi.json); POST /api/mcp speaks MCP. Do not POST /api/contact programmatically. ## Developer API - [API documentation](https://saileshdahal.com.np/developers.md): base URL, versioning policy, errors, rate limits - [OpenAPI spec](https://saileshdahal.com.np/openapi.json): typed schemas for every endpoint - [API root](https://saileshdahal.com.np/api/v1): list endpoints, read posts and search as JSON ## Optional - [Full corpus as one file](https://saileshdahal.com.np/llms-full.txt): every post concatenated, for single-fetch ingestion - [Archive](https://saileshdahal.com.np/archive.md): all posts by year, as markdown - [Home](https://saileshdahal.com.np/index.md): site overview as markdown - [Work](https://saileshdahal.com.np/work.md): shipped products, what each does and what my part was - [Contact](https://saileshdahal.com.np/contact.md): how to start a project - [JSON Feed](https://saileshdahal.com.np/feed.json): machine-readable index, each item linking its .md twin - [RSS feed](https://saileshdahal.com.np/rss.xml) Every page on this site has a markdown twin: append `.md` to any URL (`/{slug}.md`, `/archive.md`, `/series/{slug}.md`, `/tag/{slug}.md`, `/index.md`).