Official descriptionOpen source7d ago
HarshithNayakaL/HarshithNayakaL
GitHub profile of Harshith Nayaka L, an AI engineer in Bengaluru building AI agents, RAG pipelines, LLM applications and full-stack systems.
View on GitHub ↗Direct link · no redirect page
- Star count
- 0 ★ · HarshithNayakaL/HarshithNayakaL · read from the repository page · (checked 2026-10-16 20:00:20)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-10 · last seen 2026-10-10 · seen 1× · listed by 1 sources
- Score
- score 103 · 1 sources · repo
FAQ (answers come from our records)
How do I install it?
We did not find an install command for this repository (our install coverage is only about 4%). Check its README or releases page.
Who is it for?
We filed it under these tags: 开源项目. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
HarshithNayakaL/HarshithNayakaL；first catalogued 2026-10-10；last seen 2026-10-10。There is no official certification or quality rating here — we record sources and dates only.Third-party quality signals
Harshith Nayaka L
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
AI Engineer in Bengaluru (Bangalore), India. I build AI agents, RAG
pipelines and LLM applications. Currently AI Engineer at DemandNXT
(AI Engineer – Full Stack).
Most of my work lives in the gap between "the model gave an answer" and "the
system did the right thing reliably." Orchestration, structured outputs,
validation, and handling the ways models fail.
Selected work
| Project | What it is |
|---|---|
| **Octo** | Self-hosted AI research workspace built on OpenAI's Agents API, with Google Antigravity run alongside as a free-tier execution layer. Market intelligence, ICP-based sales research, continuous monitoring, cited reports and CRM integration. |
| **Maestro** · live | Multi-model orchestration engine. A conductor routes one task across thinker / worker / verifier roles and returns a verified answer with a full, replayable decision-log. |
| **BrandForge** | AI brand-to-campaign engine. Turns a brand website and a product photo into on-brand campaign creative, using Playwright, Gemini, GPT Image and n8n. |
Claude skills
| Skill | What it does |
|---|---|
| **ui-ux-taste-skill** | Discovers a user's UI taste from reference moodboards instead of vague adjectives, then locks in a stack pick and an ASCII wireframe. |
| **linkedin-algorithm-source** | LinkedIn skill built from LinkedIn Engineering's published ranking mechanism, and explicit about which popular rules are folklore. |
| **x-algorithm-source** | X strategy and writing skill derived from the open-sourced For You algorithm — real ranking weights, filter stack and credibility model. |
| **humanizer-claude-skill** | Strips AI tells from prose and scores how human it reads. --check scores 0–100 for CI, --fix rewrites. |
Also built
| Project | What it is |
|---|---|
| **CUSTOM-GPT** (Nova) | Model router. Reads query complexity and routes simple asks to cheap models, hard reasoning to stronger ones, to hold down per-chat cost. |
| **Cannon** · live | Multi-agent assistant. Each domain is an independent specialist — isolation enforced at the query, not by convention. Typed provider fallback (Groq→Gemini). 91 unit + 13 e2e tests. |
| **ReplyDesk** | WhatsApp lead agent replying in ~8s, with a live dashboard for lead feed, pipeline and response-time KPIs. |
| **CraftConnect** | Multi-modal assistant letting artisans run a storefront by talking and showing. Semi-finalist, Google Cloud Gen AI Exchange Hackathon 2025 (270,000+ developers). |
pipelines and LLM applications. Currently AI Engineer at DemandNXT
(AI Engineer – Full Stack).
Most of my work lives in the gap between "the model gave an answer" and "the
system did the right thing reliably." Orchestration, structured outputs,
validation, and handling the ways models fail.
Selected work
| Project | What it is |
|---|---|
| **Octo** | Self-hosted AI research workspace built on OpenAI's Agents API, with Google Antigravity run alongside as a free-tier execution layer. Market intelligence, ICP-based sales research, continuous monitoring, cited reports and CRM integration. |
| **Maestro** · live | Multi-model orchestration engine. A conductor routes one task across thinker / worker / verifier roles and returns a verified answer with a full, replayable decision-log. |
| **BrandForge** | AI brand-to-campaign engine. Turns a brand website and a product photo into on-brand campaign creative, using Playwright, Gemini, GPT Image and n8n. |
Claude skills
| Skill | What it does |
|---|---|
| **ui-ux-taste-skill** | Discovers a user's UI taste from reference moodboards instead of vague adjectives, then locks in a stack pick and an ASCII wireframe. |
| **linkedin-algorithm-source** | LinkedIn skill built from LinkedIn Engineering's published ranking mechanism, and explicit about which popular rules are folklore. |
| **x-algorithm-source** | X strategy and writing skill derived from the open-sourced For You algorithm — real ranking weights, filter stack and credibility model. |
| **humanizer-claude-skill** | Strips AI tells from prose and scores how human it reads. --check scores 0–100 for CI, --fix rewrites. |
Also built
| Project | What it is |
|---|---|
| **CUSTOM-GPT** (Nova) | Model router. Reads query complexity and routes simple asks to cheap models, hard reasoning to stronger ones, to hold down per-chat cost. |
| **Cannon** · live | Multi-agent assistant. Each domain is an independent specialist — isolation enforced at the query, not by convention. Typed provider fallback (Groq→Gemini). 91 unit + 13 e2e tests. |
| **ReplyDesk** | WhatsApp lead agent replying in ~8s, with a live dashboard for lead feed, pipeline and response-time KPIs. |
| **CraftConnect** | Multi-modal assistant letting artisans run a storefront by talking and showing. Semi-finalist, Google Cloud Gen AI Exchange Hackathon 2025 (270,000+ developers). |
- Source
- github · https://github.com/HarshithNayakaL/HarshithNayakaL
- Browse by field
- Open source
- added
- 7d ago
More in Open source
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.