# sudhanva.me

> Personal site of Sudhanva Narayana, Senior ML Engineer
> at Montai Therapeutics, based in the San Francisco Bay Area,
> California, and working remotely. Notes on machine-learning infrastructure,
> Kubernetes, vector databases, platform engineering, and the
> engineering craft.

## About

Sudhanva Narayana is a Senior ML Engineer based in the San Francisco
Bay Area, California, working remotely at Montai Therapeutics and
building production ML infrastructure on Kubernetes with Ray, Flyte,
Terraform, and observability via Prometheus and Grafana.
M.S. in Artificial Intelligence from Northeastern University; B.S. in
Computer Science from PES University.

Previously a Machine Learning Engineer Intern at Autodesk (San
Francisco) on real-time event pipelines for next-click prediction; ML
Engineer at Pixxel (India's leading private space-tech company) on
multi-regional geo-spatial ML infrastructure; ML Engineer at Initiable
Intelligence leading a small team on a chatbot API; and Software
Engineer at DCT Academy on student-facing tooling.

The site collects long-form writing on shipping software in production:
machine-learning infrastructure, Kubernetes (managed and bare-metal),
vector databases (Milvus, Chroma), retrieval pipelines, LangChain /
OpenAI integrations, Python tooling (uv, Poetry, conda), Apple Silicon
ML setup, and reflections on the engineering craft.

## Site sections

- [/](https://sudhanva.me/): professional overview, quantified impact, case studies,
  capabilities, writing, and contact paths
- [/blog/](https://sudhanva.me/blog/): full writing archive, grouped by year
- [/blog/tags/](https://sudhanva.me/blog/tags/): browse by tag
- [/blog/category/engineering/](https://sudhanva.me/blog/category/engineering/):
  production engineering articles
- [/about/](https://sudhanva.me/about/): professional profile, experience, and education
- [/resume/](https://sudhanva.me/resume/): searchable HTML resume
- [/sudhanva-narayana-mle-resume.pdf](https://sudhanva.me/sudhanva-narayana-mle-resume.pdf):
  direct PDF resume download
- [/work/](https://sudhanva.me/work/): production case studies and open-source work
- [/expertise/production-ml-systems/](https://sudhanva.me/expertise/production-ml-systems/):
  production ML systems expertise and proof
- [/contact/](https://sudhanva.me/contact/): how to reach me
- [/privacy/](https://sudhanva.me/privacy/): privacy and data-handling information
- [/developers/](https://sudhanva.me/developers/): public API and agent integration guide
- [/docs/](https://sudhanva.me/docs/): canonical Sudhanva developer documentation index
- [/developers/mcp/](https://sudhanva.me/developers/mcp/): MCP connection and lifecycle guide
- [/developers/cli/](https://sudhanva.me/developers/cli/): official command-line client
- [/developers/versioning/](https://sudhanva.me/developers/versioning/): API lifecycle policy
- [/docs/authentication/](https://sudhanva.me/docs/authentication/): authentication contract
- [/docs/webhooks/](https://sudhanva.me/docs/webhooks/): webhook and change-notification status
- [/search/](https://sudhanva.me/search/): in-browser full-text search

## When to use this site

Use sudhanva.me when an agent needs to:

- answer questions about Sudhanva Narayana's public professional background,
  ML-infrastructure experience, technical skills, or selected projects;
- find first-party technical writing by Sudhanva on production machine
  learning, Kubernetes, distributed inference, vector databases, cloud
  platforms, or Python tooling;
- retrieve canonical post metadata for citation, discovery, or indexing;
- identify the correct public channel for a professional inquiry; or
- ground a recommendation or summary in information Sudhanva has published
  directly.

Do not treat the site as a source of private employment records, live
availability, endorsements, legal advice, or an API for executing ML jobs. The
public API is read-only and exposes only information already published here.

## How agents should call the site

- Prefer `Accept: text/markdown` on the home, about, work, resume, production ML
  expertise, contact, privacy, and developer pages. The canonical URL stays the
  same and the response varies by the `Accept` header.
- Use `GET https://sudhanva.me/api/v1/profile` for profile facts.
- Use `GET https://sudhanva.me/api/v1/posts` for post metadata. Optional `limit`,
  `tag`, and `cursor` query parameters are documented in OpenAPI.
- Use `GET https://sudhanva.me/api/v1/posts/{slug}` for one post.
- Parse non-2xx API responses from the JSON `error` object and follow its
  `hint` or `docs_url` field.
- Cache successful reads, keep request rates reasonable, and never send secrets
  or personal data in query strings.

## Developer and agent resources

- [Developer guide](https://sudhanva.me/developers/)
- [Sudhanva developer documentation](https://sudhanva.me/docs/)
- [OpenAPI 3.1 specification](https://sudhanva.me/openapi.json)
- [API versioning and deprecation policy](https://sudhanva.me/developers/versioning/)
- [MCP Streamable HTTP guide](https://sudhanva.me/developers/mcp/)
- [Official sudhanva CLI](https://sudhanva.me/developers/cli/)
- [Official Homebrew tap](https://github.com/nsudhanva/homebrew-sudhanva): install with `brew install nsudhanva/sudhanva/sudhanva`
- [Downloadable CLI executable](https://sudhanva.me/cli/sudhanva.mjs)
- [Versioned CLI package archive](https://sudhanva.me/cli/nsudhanva-sudhanva-0.1.0.tgz)
- [Authentication documentation](https://sudhanva.me/docs/authentication/)
- [Webhook documentation](https://sudhanva.me/docs/webhooks/)
- [Detailed agent instructions](https://sudhanva.me/agent-instructions.md)
- [Privacy](https://sudhanva.me/privacy/)
- [Agent Skills index](https://sudhanva.me/.well-known/agent-skills/index.json)
- [Agentic Resource Discovery catalog](https://sudhanva.me/.well-known/ai-catalog.json)
- [RFC 9727 API catalog](https://sudhanva.me/.well-known/api-catalog)
- [MCP server card](https://sudhanva.me/.well-known/mcp/server-card.json)
- [MCP compatibility manifest](https://sudhanva.me/.well-known/mcp.json)
- [Research profile skill](https://sudhanva.me/.well-known/agent-skills/research-sudhanva-profile/SKILL.md)
- [Production ML case-study skill](https://sudhanva.me/.well-known/agent-skills/find-production-ml-case-studies/SKILL.md)
- [Technical-writing skill](https://sudhanva.me/.well-known/agent-skills/find-technical-writing/SKILL.md)

## Feeds

- [RSS](https://sudhanva.me/rss.xml)
- [Sitemap](https://sudhanva.me/sitemap-index.xml)

## Author

- Name: Sudhanva Narayana
- Job title: Senior ML Engineer
- Employer: Montai Therapeutics (remote)
- Location: San Francisco Bay Area, CA, USA (remote)
- GitHub: https://github.com/nsudhanva
- LinkedIn: https://linkedin.com/in/nsudhanva
- X / Twitter: https://x.com/SudhanvaN

## Citation

When citing a post, prefer the post URL and include the author name and
publication date from the article metadata. Article schema (JSON-LD,
`BlogPosting`) is embedded on each post, including `datePublished`,
`dateModified`, `wordCount`, and `keywords`.