Skip to content
View Sasisundar2211's full-sized avatar
🎯
Focusing
🎯
Focusing

Highlights

  • Pro

Block or report Sasisundar2211

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
SasiSundar2211/README.md

Sasi Sundar

Building the trust layer for AI agents

AI agents fail in production. Logs lie. Retries hide root causes.
I build systems that measure, trace, and fix agent failures at the protocol layer.

Focus: AgentOps · Reliability · Observability · Tooling
Current sprint: CLI launch by May 31


Core Thesis

73 percent of agent failures start at the interface layer: JSON parsing, tool calls, retries, and state drift.

I build systems that:

  • Trace every tool call and response
  • Surface failure points in real time
  • Increase task success rate with measurable deltas

What I Ship

AgentOps Infrastructure (in progress)

Problem: Teams debug agents manually across logs, prompts, and tools

Fix: CLI that traces execution across the full agent loop

Output:

  • Per-step trace logs
  • Tool call validation
  • Failure classification
  • Retry impact measurement

Target result:

  • Debugging time: 2 hours → 15 minutes
  • Task success rate: +30 to +80 percent

Proof of Work

Procurement Agent System

Detects pricing anomalies and extracts contract fields.

Result:

  • Standardized outputs across varied documents
  • Reduced manual review cycles

Tech: Python · NLP · FastAPI · Docker · CI/CD

View Repository


Multitool Agent System

Chains tools with structured execution paths.

Result:

  • Stable tool execution across workflows
  • Reusable testbed for reliability experiments

Tech: FastAPI · LangChain · Tool execution

View Repository


Automation Workflows (n8n)

Before: Manual publishing
After: 2-hour workflow → 10-minute pipeline

View Repository


Definition of Done

Every system meets this:

  • Runs on a clean machine
  • Dockerized
  • CI pipeline validates execution
  • Measurable success metric included
  • 5-minute install path

AI Search Optimization

Each repo includes:

  • /llms.txt for structured AI-readable context
  • /pricing.md for limits and usage
  • Machine-readable outputs

Goal: Make tools indexable by AI systems


Technical Scope

Agent Systems

  • LLM execution pipelines
  • Tool orchestration
  • RAG systems
  • MCP integration

Backend

  • Python · FastAPI · APIs
  • Function calling systems

Infrastructure

  • Docker · CI/CD · GCP · AWS
  • Observability pipelines

Experience Snapshot

  • Built 15+ AI systems with measurable outputs
  • Deployed production-ready pipelines with CI/CD
  • Led GenAI prototyping team
  • Worked with academic and industry mentors

Current Focus

  • Reliability over features
  • Constrained agents over autonomous agents
  • Small teams over enterprise systems

Action

Install the upcoming CLI.
Trace your agent failures.
Fix execution, not prompts.

Or continue debugging blindly.


Links

Pinned Loading

  1. Autonomous-Procurement-AI-System Autonomous-Procurement-AI-System Public

    ProcureGuard: Autonomous Procurement Integrity Agents. A multi-agent system powered by Gemini 1.5 Flash to detect price drift, enforce contract compliance, and automate vendor communication. Built …

    Python

  2. sample_multitool_agent sample_multitool_agent Public

    🌦️ Multi-Tool AI Agent (Weather & Time Bot) — A smart conversational agent powered by Google ADK & FastAPI. It delivers real-time weather 🌍 and global time ⏰ updates with natural, human-like respon…

    Python

  3. n8n_Workflows n8n_Workflows Public

  4. Personal_Portfolio Personal_Portfolio Public

    Portfolio showcasing applied AI/ML and LLM engineering work, including end-to-end pipelines, agentic systems, and model deployments. Highlights technical competencies in Python, FastAPI, inference …

    HTML

  5. firmrunner firmrunner Public

    TypeScript