Technical Writer – AI/ML Deep-Dive Article (SkillOpt vs GEPA vs DSPy)
Budget: $2 – $8 USD
We need a skilled technical writer to produce a single, well-researched article comparing SkillOpt, GEPA, and DSPy — three frameworks relevant to AI agent skill optimization and prompt programming. This is a paid trial engagement with potential for ongoing work.
The article must match a specific first-person operator voice: precise, sourced, no hype, no generic AI-blog filler. You will be provided a source packet for the topic and must write from that packet plus publicly cited sources only. Do not invent internal architecture not provided.
Requirements:
• Strong background writing about AI/ML systems, agent frameworks, or developer tooling
• Ability to write in a tight, first-person technical voice — not thought-leadership fluff
• Comfortable citing sources explicitly and marking unknowns as TBU rather than guessing
• Experience producing structured long-form technical content (H2/H3 outlines, source lists)
• Familiarity with prompt optimization frameworks (DSPy knowledge is a strong plus)
• Ability to hit a 4–6 hour cap and communicate blockers early
Deliverables:
• H2/H3 outline submitted within 24 hours of kickoff
• Full draft: 1,200–1,800 words in Google Doc or Markdown
• Source list with links and a note on what each source proved
• 3 title options + 1 meta description (155 characters max)
• Optional: rough HTML fragment for one diagram or code block
• One revision round included
Work arrangement: Fixed-price trial. If the trial is successful, ongoing work is hourly at 3–5 hours per article depending on complexity. Draft-only — no publishing access required.
About the project: This is a technically sophisticated personal blog covering AI workflows, FP&A, and operator-level tooling. The audience is technical and expects depth, precision, and honest analysis. Upcoming topics include recursive self-improvement with AI, intelligent model routing, and LangChain/LangFuse tracing — strong candidates will be excited by this stack.
The article must match a specific first-person operator voice: precise, sourced, no hype, no generic AI-blog filler. You will be provided a source packet for the topic and must write from that packet plus publicly cited sources only. Do not invent internal architecture not provided.
Requirements:
• Strong background writing about AI/ML systems, agent frameworks, or developer tooling
• Ability to write in a tight, first-person technical voice — not thought-leadership fluff
• Comfortable citing sources explicitly and marking unknowns as TBU rather than guessing
• Experience producing structured long-form technical content (H2/H3 outlines, source lists)
• Familiarity with prompt optimization frameworks (DSPy knowledge is a strong plus)
• Ability to hit a 4–6 hour cap and communicate blockers early
Deliverables:
• H2/H3 outline submitted within 24 hours of kickoff
• Full draft: 1,200–1,800 words in Google Doc or Markdown
• Source list with links and a note on what each source proved
• 3 title options + 1 meta description (155 characters max)
• Optional: rough HTML fragment for one diagram or code block
• One revision round included
Work arrangement: Fixed-price trial. If the trial is successful, ongoing work is hourly at 3–5 hours per article depending on complexity. Draft-only — no publishing access required.
About the project: This is a technically sophisticated personal blog covering AI workflows, FP&A, and operator-level tooling. The audience is technical and expects depth, precision, and honest analysis. Upcoming topics include recursive self-improvement with AI, intelligent model routing, and LangChain/LangFuse tracing — strong candidates will be excited by this stack.