RECRUITER LENS // CANDIDATE EVALUATION DOSSIER

Role Fit Evaluation // Evidence Matcher

Evaluate Arun Jyoti Chakraborty against your specific open requirements. Transparent evidence match scores, direct repository proofs, and 1-click ATS summary notes formatted for hiring managers.

Select Hiring Focus Lens:

EXCEPTIONAL FITSCORE: 96 / 100

Agentic AI Systems Engineer

Autonomous ReAct Loops · FastMCP Tool Servers · Multi-Agent Orchestration · Safety Guardrails

Arun demonstrates hands-on mastery of decoupled multi-agent architectures using Google ADK and LangChain, implementing stateful ReAct loops, Model Context Protocol (MCP) tool sandboxing, and deterministic safety test suites.

Evidence Match

96%

Verified Proofs

01. "Why Hire Arun?" // Business ROI & Tangible Impact

FOR HIRING MANAGERS

✓

Immediate Day-1 productivity in architecting tool-calling LLM agents (FastMCP, LangChain @tool decorators).

✓

Eliminates retail financial advice risk and adversarial prompt injection via automated pytest guardrail suites (100% pass rate).

✓

Engineers decoupled HTTP/SSE tool layers rather than monolithic agent spaghetti code, ensuring clean separation of concerns.

✓

Hands-on experience deploying stateful session persistence (Google ADK SqliteSessionService).

02. Role Requirement to Repository Proof Matrix

EMPIRICAL BENCHMARKS

Multi-Agent Orchestration & Planningproduction

Built Google ADK agent with automated fallback chains (Gemini 1.5 Pro -> Gemini 2.5 Flash -> Flash-8b).

Repo Proof: Algorithmic Trading Agent (adk-orchestrator)

Tool Sandboxing & Context Protocolsverified

Engineered independent FastMCP server exposing live market data and pmdarima forecasting tools over HTTP/SSE.

Repo Proof: FastMCP Tool Server (FastMCP v0.4)

Safety & Adversarial Jailbreak Defenseverified

Automated pytest suite testing 50+ boundary injections to guarantee deterministic refusal of retail financial advice.

Repo Proof: Pytest Guardrail Suite (test_guardrails.py)

Autonomous ReAct Loop Executionproduction

LangChain ReAct agent capable of multi-step financial research (price fetch -> income comparison -> trend synthesis).

Repo Proof: AI Financial Analyst Agent (react-planner)

03. Verified Production Systems Backing This Role

INTERACTIVE INSPECTOR

agent→

Algorithmic Trading & Predictive Research Agent

Decoupled multi-agent quantitative framework with FastMCP and ARIMA forecasting.

agent→

AI Financial Analyst Agent

Autonomous market research agent executing ReAct multi-step planning and headless chart generation.

agent→

Healthcare Appointment Assistant (MyEyeDr)

Conversational healthcare assistant with patient classification and Salesforce CRM automation.

04. 1-Click ATS Candidate Summary for Hiring Manager / Slack

PRE-FORMATTED NOTES

CANDIDATE: Arun Jyoti Chakraborty (IIT Guwahati, B.Sc. Data Science & AI)
TARGET ROLE: Agentic AI Systems Engineer (Match: 96% - Exceptional Fit)
KEY HIGHLIGHTS:
- Capstone Project for Google & Kaggle 5-Day AI Agents Intensive.
- Built decoupled multi-agent systems using Google ADK and FastMCP tool servers over HTTP/SSE.
- 100% Pass Rate on automated pytest guardrails preventing prompt injection and unvetted advice.
- Production containerized deployments on AWS EC2 behind Cloudflare SSL.
LIVE EVIDENCE: https://arunjyoticode.me/systems/algorithmic-trading-agent