Skip to content
View adiseshvsanklapur's full-sized avatar
🏁
Learning Fast, Collaborating
🏁
Learning Fast, Collaborating
3 followers · 1 following
  • Pleasanton, CA

Achievements

Achievement: Pull Sharkx2Achievement: Arctic Code Vault Contributor

Achievements

Achievement: Pull Sharkx2Achievement: Arctic Code Vault Contributor

Block or report adiseshvsanklapur

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
adiseshvsanklapur/README.md
Adisesh Sanklapur. Software engineer and ML researcher. Now: Software Engineering Intern at Shopify, search index serving. Published: ICML 2026, first author. Education: UC Davis, B.S. Computer Science, Dec 2027.

Portfolio  ·  LinkedIn  ·  Email

Featured work

Cache-Aware Inference Router: a C++20 reverse proxy that pins repeat prompts to the vLLM replica already holding their KV cache. −43% p99 latency, 79% fleet cache hits, under 0.9 ms routing per request, zero divergence over 50K generations. Limit Order Book and Matching Engine in C++20: 23M messages per second, 6.1x faster cancel than std::map, zero divergences across 30M fuzzed messages.
ICML 2026, first author: On the Transfer of Output Diversity via Synthetic Data in Language Models. Light tuning kept 29–39% of diversity gains; heavy tuning collapsed to 68.4% top-5 repetition; 1,250 generations evaluated. AccessMap AI, Top 3 of 100+ teams at HackDavis 2026: wheelchair-accessible routing over a 19.7K-node graph, under 300 ms per query.
Hypothesis-Driven Equities, Top 5 at Cornell 2023: a systematic trading research platform. 3 hypotheses tested, 1 validated, 2 rejected. LLM Inspector: hypothesis-driven LLM evaluation with LLM-as-judge, SBERT similarity, and RAG faithfulness checks, running locally on Ollama.

Experience

  • Shopify: Software Engineering Intern, search index serving
  • ARMS Lab, UC Davis: AI Systems Researcher
  • Pinpoint, Codelab Davis: Software Developer
  • Arcadia: Software Engineering Intern, robotics

Toolkit

Languages  C++17/20 · Python · SQL · TypeScript · Go · Java · C#
Systems & ML  vLLM · PyTorch · LoRA/PEFT · MLX · FastAPI · PostgreSQL · Redis · Kafka
Infra  Docker · Kubernetes · AWS · GCP · Linux · CI/CD

Pinned Loading

  1. High-Performance-LOB High-Performance-LOB Public

    C++

  2. Hypothesis-Driven-Equities Hypothesis-Driven-Equities Public

    Python

  3. LLM-Inspector LLM-Inspector Public

    JavaScript

  4. ai-slides-generator ai-slides-generator Public

    Python

  5. portfolio portfolio Public

    CSS

  6. PhotoEditor PhotoEditor Public

    A photo editor to help you edit your photos using quality tools!

    Java 1