Skip to content
View RudraDudhat2509's full-sized avatar
🏠
Working from home
🏠
Working from home
12 followers · 15 following

Achievements

Achievement: Pair Extraordinairex3Achievement: QuickdrawAchievement: YOLOAchievement: Pull Shark

Achievements

Achievement: Pair Extraordinairex3Achievement: QuickdrawAchievement: YOLOAchievement: Pull Shark

Block or report RudraDudhat2509

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.

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
RudraDudhat2509/README.md
Rudra Dudhat

I build AI agents, then I break them. Professionally.

Production AI · agents · automations · the 2am debugging nobody posts about.

IIT Bhilai CGPA Open

Portfolio Field Notes LinkedIn X Email

Shipped upstream

Merged PRs into other people's production codebases: observability, AI gateways, eval tooling, security scanners. This whole block updates itself daily, because hardcoding your own PRs is embarrassing.

Repo Contribution Merged
Arize-ai/phoenix fix: preserve generator inputs in create_dataset and add_examples_to_dataset 2026-06-30
open-telemetry/opentelemetry-python opentelemetry-sdk: drop NaN measurements at instrument level to prevent aggregation poisoning 2026-06-25
kedro-org/kedro fix: reject Python keywords as pipeline names in kedro pipeline create 2026-06-22
kedro-org/kedro fix(cli): reject project names that shadow stdlib modules or keywords 2026-06-19
usestrix/strix fix: route ollama models through ollama_chat so tool calling works 2026-06-16
BerriAI/litellm fix: evict last deleted model in multi-instance deployments 2026-06-12
mlflow/mlflow fix: extend mlflow.sourceRun metrics filter to cover post-hoc linked OTLP traces 2026-05-27
AaravWayneBanner/3GPP_Channel_Model docs: add comprehensive codebase documentation 2026-05-23
firstcontributions/first-contributions Add Rudra Dudhat to Contributors list 2026-02-20
ed-donner/llm_engineering Add Market Research Agent notebook to community-contributions 2025-08-23

🛠 Selected work

Project The problem What shipped
diffprompt prompt changes ship without tests LLM-as-judge cascades + HDBSCAN drift detection · live on PyPI
Cascade AI multi-agent backends die quietly self-healing routing + per-user Firestore memory, live on Firebase
OptiQuant black-box trading models 28.16% CAGR · 1.05 Sharpe · SHAP-explained · deployed, not abandoned in a notebook
Outreach Engine outreach that reads like a bot LangGraph pipeline with a human approval gate in Gmail
@ Altagic under NDA in production, running daily. that's all you get

🧰 Stack

Python FastAPI LangGraph OpenTelemetry PyTorch Docker Postgres AWS

📓 The rest

Working notes that publish themselves straight from a repo: Field Notes (push a .txt, it shows up typeset). Built-for-the-plot side quests: ragbait, iceman, ctxbridge.

Peer reviewed: funny · delusional · ambitious. Fueled by tea, not coffee.

Pinned Loading

  1. diffprompt diffprompt Public

    git diff for prompt engineers

    Python 2

  2. RudraDudhat2509.github.io RudraDudhat2509.github.io Public

    portfolio-website

    HTML

  3. cacheguard cacheguard Public

    Drop-in safety layer that stops semantic caches from silently returning wrong answers — with the first independent open benchmark of semantic-cache false-hit rates. Free, local, LiteLLM + OpenTelem…

    Python 1

  4. brok brok Public

    An MCP that reviews your architecture, estimates real capacity from cited numbers, surfaces the trade-offs you are making, and roasts bad design in a dry voice. Deterministic engine, no model, runs…

    Python 1

  5. mimic-eval mimic-eval Public

    Distill expensive LLM judges into cheap, auditable deterministic code. Always shadows the real judge before it takes over.

    Python

  6. receipts receipts Public

    Receipts: an MCP server that runs public-data due diligence on any brand and returns a verdict-first report with receipts. Deterministic scoring, abstention over hallucination.

    Python