Skip to content
View mbsdeepak's full-sized avatar
1 follower · 0 following

Block or report mbsdeepak

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
mbsdeepak/README.md
Bhargava Sai Deepak Mathukumalli — building LLM-agent infrastructure from first principles

I build cloud cost-intelligence and data platforms professionally, and implement LLM-agent infrastructure from the ground up on my own time — one small, single-purpose library per hard problem, no framework in the way. The bias throughout: build the thing to understand it — minimal dependencies, code you can read end-to-end, behaviour you can inspect.


⌁   An agent platform, built in layers

Each hard part of running an LLM agent, implemented as an independent library. They stand alone — and compose into a complete platform.

flowchart LR
  T(["request"]) --> BH["🛡️ bulkhead<br/><i>gateway</i>"]
  BH --> CG["⚙️ cogs<br/><i>runtime</i>"]
  LM["🧵 loom<br/><i>context</i>"] --> CG
  CG --> SN["📡 sonar<br/><i>tracing</i>"]
  CG -. evals .-> GT["🎯 gauntlet"]
  classDef n fill:#161b22,stroke:#30363d,color:#e6edf3,rx:8,ry:8;
  class T,BH,CG,LM,SN,GT n;
Loading
Project What it solves Notable internals
cogs The agent runtime Agent loop, typed tool protocol, provider abstraction, context management, permissions, and deterministic record/replay — in ~1.5k lines.
loom Context engineering Chunking, embeddings, vector retrieval, history compaction, and token-budgeted context assembly.
bulkhead Provider resilience Retries, circuit breaking, rate limiting, caching, failover, and cost governance in front of any provider.
gauntlet Agentic evaluation Deterministic simulated tool environments, state/trajectory/LLM-judge grading, and pass@k with Wilson confidence intervals.
sonar Run observability OpenTelemetry-style span tracer, cost/latency meters, and text/HTML trace timelines.
qwery SQL over raw files Hand-written tokenizer + recursive-descent parser feeding a Volcano-model executor over CSV/Parquet — the design behind DuckDB, in pure Python.

⌁   Currently

Cloud cost-intelligence — an LLM agent that answers cost questions over large-scale usage data, and the analytics pipelines behind it on Kubernetes.

Reach for:   Python   Go   C++   AWS · Bedrock · Athena   Kubernetes · Argo   SQL


$ contact  ·  LinkedIn  ·  mbsdeepak3@gmail.com

Pinned Loading

  1. qwery qwery Public

    A tiny SQL query engine that runs SELECT queries directly over CSV and Parquet files — pure-Python, no database, no server. Hand-written tokenizer and recursive-descent parser feed a Volcano-model …

    Python 1

  2. bulkhead bulkhead Public

    A resilient LLM gateway — retries, circuit breaking, rate limiting, caching, failover, and cost governance in front of any provider, in ~2k readable lines of Python.

    Python

  3. cogs cogs Public

    A minimal but real LLM agent runtime — the core of a coding-agent harness (agent loop, typed tool protocol, provider abstraction, context management, permissions, and deterministic record/replay) i…

    Python

  4. gauntlet gauntlet Public

    A rigorous agentic tool-use eval set for LLMs: deterministic simulated tool environments, state/trajectory/LLM-judge grading, pass@k with Wilson confidence intervals, and cost/latency tracking.

    Python

  5. loom loom Public

    The context-engineering layer for an agent — chunking, embeddings, vector retrieval, history compaction, and token-budgeted context assembly, in ~2k readable lines of Python.

    Python

  6. sonar sonar Public

    Observability for agent runs — an OpenTelemetry-style span tracer, cost/latency meters, trace ingest for cogs & gauntlet, and text/HTML timelines, in ~2k readable lines of Python.

    Python