asopi tech
asopi techIndie Developer

Git / Blog evidence

About AsopiTech

AsopiTech is a public software product R&D studio. We publish new software designs as code, data, comparisons, and articles, then grow them into open-source projects and products. Akari Takamine guides from the market side; Asmati Tatsumi explains from the technical side.

Profile

Profile

Over 20 years of software development experience

With over 20 years in software development, I have hands-on experience spanning cloud infrastructure, frontend technologies, blockchain, and AI.

  • Development experience with Java / C# / Python / Go / TypeScript
  • Production experience with Kubernetes / Docker / k3s / GKE
  • Frontend development with TypeScript / Angular / React / Astro
  • Blockchain development with Hyperledger Fabric / Indy
  • Cloud infrastructure on AWS / GCP / Azure
  • AI utilization with ChatGPT / Auto-GPT / Stable Diffusion
  • IoT projects with Raspberry Pi / M5Stack
  • OSS product development in Rust / Nim (Alopex DB, Nimino / Nimculus)

Media

Blog Posts

Technical notes and development logs

All posts
  • Blog

    Indexes and the Art of Searching, Part 8: Choosing Methods and a Decision Tree

    Published methods are tuned to the data they were evaluated on. How should you budget and plan the work of tuning them to your own data instead?

    Read post
  • Blog

    Indexes and the Art of Searching, Part 7: Evaluation Functions That Drive Search

    Search only moves once something tells it how good a candidate is. How are heuristics, game evaluations, and verifiers built, and where do they break?

    Read post
  • Blog

    Indexes and the Art of Searching, Part 6: Local Search and Nature-Inspired Methods

    Sometimes getting a usable solution quickly matters more than finding the optimal one. Which components do local search and animal-inspired methods let you tune?

    Read post
  • Blog

    Indexes and the Art of Searching, Part 5: State-Space Search

    Car navigation routes and next moves in shogi have too many candidates to survey. What does a search that alternates local evaluation and global pruning prepare in advance?

    Read post
  • Blog

    Indexes and the Art of Searching, Part 4: Similarity Search and Recommendation

    When neither the keys nor the words match, how do you find things that are close? This part sorts out how search and recommendation pull candidates by distance and order them by behavior.

    Read post
  • Blog

    Indexes and the Art of Searching, Part 3: Searching by Words and Ranking Results

    Once you find the documents that contain a word, you still have to decide which to read first. After an inverted index narrows the candidates, what should the ranking rest on?

    Read post