RC
Rakesh ChaudharyEx-Agentic AI @ BEL
Ex-Agentic AI Intern @ Bharat Electronics Limited (BEL)DTU '27

Architecting Multi-Agent RAG & Scalable Systems

Hi, I'm Rakesh Chaudhary. Final Year B.Tech student in Mathematics & Computing at DTU and former Agentic AI Intern at BEL. I engineer retrieval-augmented generation pipelines, multi-agent workflows, and high-performance backend systems.

import { Developer } from '@dtu/maths-computing';

import { AgenticAI, RAG } from '@bel/central-research-lab';


export const rakesh: Developer = {

name: 'Rakesh Chaudhary',

education: {

institution: 'Delhi Technological University (DTU)',

degree: 'B.Tech Mathematics & Computing',

cpi: 6.4,

duration: '2023 - 2027'

},

experience: {

role: 'Ex-Agentic AI Intern',

organization: 'Bharat Electronics Limited (BEL)',

period: 'June 2026 - July 2026',

focus: ['Multi-Agent RAG', 'Vector Databases', 'Defense Doc QA']

},

problemSolving: {

dsaProblemsSolved: 600,

codeforcesPeakRating: 1620 // Expert,

codeforcesHandle: 'rakeshsiwal'

},

links: {

linkedin: 'linkedin.com/in/rakesh-chaudhary--',

github: 'github.com/rakesh404e'

}

};

01. CAREER EXPERIENCE

Work Experience

Enterprise Defense AI & Agentic Systems

Agentic AI Intern

Completed Internship

Bharat Electronics Limited (BEL) — Central Research Laboratory - BEL

June 2026 – July 2026
Ghaziabad, UP
  • ›Built enterprise Retrieval-Augmented Generation (RAG) microservice querying defense records, indexing 25,000+ docs in ChromaDB & FAISS.
  • ›Engineered ingestion pipeline (parsing → chunking → embedding), cutting manual lookup time by 85%.
  • ›Designed multi-agent reasoning workflows using LangGraph & LangChain for automated retrieval, boosting precision by 34%.
  • ›Implemented hybrid search, cross-encoder reranking, and RBAC, achieving 92% top-k accuracy with zero data leakage.
  • ›Shipped containerized LLM deployment using Docker & FastAPI, monitored with structured telemetry.
  • ›Integrated modular AI agents, retrieval tools, and model-serving components into production-oriented workflows, enabling scalable and independently executable AI tasks.
PythonLangGraphLangChainRAGChromaDBFAISSFastAPIDockerHybrid SearchRBAC
02. FEATURED ENGINEERING

Featured Projects

AI Pipelines, Multi-Agent Systems & Computer Vision

Agentic AI & RAG2026

AI Agent Chatbot — Memory & Retrieval System

Python, FastAPI, SQLite, ChromaDB, Docker, Nginx

Engineered hybrid retrieval combining ChromaDB semantic search and SQLite keyword matching via Reciprocal Rank Fusion (RRF), achieving 81.67% Recall@3 and 100% user isolation across 60-query benchmark.

✓Engineered hybrid retrieval combining ChromaDB semantic search and SQLite keyword matching via RRF, achieving 81.67% Recall@3 and 100% user isolation across 60-query benchmark.
✓Reduced user latency by 30.21% (12.22s → 8.53s P50) via async memory extraction and vector indexing with 0 background failures.
✓Reduced LLM context size by 98.73% on 30-message histories using rolling conversation summarization.
✓Architected synchronized SQLite + ChromaDB memory with user caching (90% hit rate), achieving 100% update/reversal success with 0 duplicate keys.
✓Added fault injection, achieving 100% pass rate across 5 reliability scenarios and 43/43 passing tests.
81.67%
Recall@3
-30.21%
Latency P50
98.73%
Context Shrink
PythonFastAPISQLiteChromaDBDockerNginxRRF Hybrid SearchRolling Summary
Multi-Agent Systems2026

Multi-Agent AI Engineering Platform

AWS EC2, Python, LangGraph, MCP, Docker, GitHub, OmniRoute, AWS SSM

Architected self-hosted multi-agent platform on AWS EC2, coordinating 10+ specialized agents via hierarchical task decomposition and parallel execution.

✓Architected self-hosted multi-agent platform on AWS EC2, coordinating 10+ specialized agents via hierarchical task decomposition and parallel execution.
✓Built orchestration layer using LangGraph to decompose objectives and track state; integrated OmniRoute gateway with fallback routing across LLM providers.
✓Integrated MCP + AWS SSM tool layer for secure command execution; containerized runtime with Docker on 2-vCPU / 7.6-GB RAM EC2 instance.
✓Implemented GitHub-based workflows enabling agents to edit code, manage branches, run tests, and produce structured deliverables.
✓Designed modular architecture separating orchestration, agents, tools, skills, and policies for concurrent execution and expansion.
10+ Coordinated
Agents
2-vCPU / 7.6GB
Runtime Spec
LangGraph
Orchestration
AWS EC2PythonLangGraphMCPDockerGitHub WorkflowsOmniRouteAWS SSM
Serverless AI Platform2026

Test-Case Triage Coach

AWS Lambda, Bedrock, LangGraph, DynamoDB, API Gateway, S3, CloudFront, AWS SAM, Astro, TypeScript, Docker

Built serverless AI debugging platform using AWS Lambda, API Gateway, DynamoDB, LangGraph, & Bedrock for progressive 3-tier DSA hints.

✓Built serverless AI debugging platform using AWS Lambda, API Gateway, DynamoDB, LangGraph, & Bedrock for progressive 3-tier DSA hints.
✓Developed Python code-execution sandbox with subprocess isolation, restricted modules, and 2s execution timeout.
✓Automated deployment via AWS SAM & GitHub Actions with 45+ pytest/moto tests; hosted frontend on S3 + CloudFront.
Progressive 3-Tier
Hint System
2s Isolated
Sandbox Timeout
45+ Pytest
Automated Tests
AWS LambdaAWS BedrockLangGraphDynamoDBAWS SAMAstroTypeScriptDocker
Computer Vision & AI2025

Traffic Flow Optimization System

Python, YOLOv8, OpenCV, Deep Learning

Developed a real-time computer vision system using YOLOv8 for vehicle detection and queue counting at 45+ FPS with adaptive traffic light scheduling.

✓Developed real-time computer vision system using YOLOv8 for vehicle detection and queue counting at 45+ FPS.
✓Designed adaptive density estimation dynamically optimizing signal green-times, reducing wait times by 32%.
✓Built video processing pipelines to monitor multi-lane congestion and log traffic analytics.
45+ FPS
Processing
-32%
Wait Time
YOLOv8
Architecture
PythonYOLOv8OpenCVDeep LearningReal-Time Video
Quantitative Finance2024

AlgoRisk Insights

Finance and Analytics Club

Architected quantitative trading strategies and automated backtesting engines incorporating statistical cointegration, risk modeling, and trailing risk management.

✓Built algorithmic strategies like Pairs Trading and Stop Loss using Mean Reversion logic and technical indicators.
✓Enhanced returns and risk metrics via denoising and statistical tests (ADF, KPSS, Ljung-Box, Granger Causality).
✓Built Backtesting Engine incorporating Trailing Stop Loss, Take Profit, CAPM, and CPPI.
Mean Reversion
Logic
CAPM / CPPI
Risk Engine
ADF / KPSS
Tests
PythonQuantitative FinancePairs TradingBacktestingTime SeriesRisk Engines
03. TECHNICAL CAPABILITIES

Skills & Technologies

AI Frameworks, Distributed Systems & Core CS

⚡

AI, LLM & Agentic AI

LangGraphLangChainRAGMCPTool CallingMulti-Agent WorkflowsLLM APIsPrompt EngineeringHugging Face
💻

Programming Languages

PythonC/C++JavaScriptSQLBash
📊

Data & ML Stack

ChromaDBFAISSHugging Face EmbeddingsPandasNumPyScikit-learnOpenCV
⚙️

Frameworks & Web

FastAPIFlaskReactNode.js
☁️

Cloud & DevOps

AWSGCPDockerGitLinux/Unix
🧠

Core CS & Math

DSADBMSOperating SystemsComputer NetworksProbability & Statistics
04. RECOGNITION & CODING

Achievements & Milestones

Competitive Programming & Academic Honors

🏆
1620+
Codeforces Peak Rating
Expert tier on platform (handle: rakeshsiwal)
🧩
600+
DSA Problems Solved
Across LeetCode and Codeforces platforms
🎯
93.25%
JEE Main Percentile
Conducted by NTA among 1.2M candidates

Programming Achievements

LeetCode & Codeforces

Codeforces Expert (Peak 1620+)

Codeforces

Achieved Expert tier rating on Codeforces under handle rakeshsiwal.

View Codeforces Profile (rakeshsiwal)

Global Ranks 466 & 312 in Div. 2

Global Contests

Secured top international ranks in Codeforces Rounds (Div. 2) among 15,000+ global participants.

600+ Algorithmic Problems Solved

DSA

In-depth problem solving in Data Structures, Graph Theory, Dynamic Programming, and Math across LeetCode & Codeforces.

Top 15% in AlgoUniversity Fellowship

Fellowship '24

Ranked among top 15% in AlgoUniversity Technology Fellowship 2024 Stage 1 out of 20,000+ candidates.

Scholastic Achievements

National Exams & Honors

93.25 Percentile in JEE Mains 2023

JEE Main

Conducted by the National Testing Agency (NTA) among 1.2 million nationwide candidates.

Qualified for JEE Advanced 2023

JEE Advanced

Conducted by IIT Guwahati among 180,000+ shortlisted candidates across India.

Academic Excellence Award

CBSE Board

Awarded for top 1% scholastic performance in High School Board Examinations.

Leadership, Creative & Community Activities

DTU Leadership & Campus Roles

Cultural Council, DTU — Head

Leadership

Head of the Cultural Council at Delhi Technological University, contributing to the organization and execution of major student cultural activities including Engifest, DTU's flagship cultural festival.

Mathics Society, DTU — Active Member

Academic Club

Active member of the Mathics Society, participating in mathematics-focused student activities, seminars, and peer initiatives.

Events & Hackathons

Organizing

Involved in organizing multiple college events and hackathons, supporting logistics, team coordination, and event execution.

Art & Creative Activities

Creative Honors

Received multiple awards and recognitions in art and craft during school, with continued involvement in sketching, painting, and creative work.

Social & Community Engagement

Community

Participated in NGO-led and community-oriented social improvement activities.

05. ACADEMICS

Education & Coursework

Academic Credentials & Core Curriculum

Final Year Undergrad2023 – Present

Delhi Technological University (DTU)

B.Tech (Mathematics & Computing)

CPI: 6.4 / 10

Department of Mathematics and Computing

  • ›Rigorous program blending Advanced Pure & Applied Mathematics, Computer Science, and Deep Learning.
  • ›Specialized in Agentic AI, Multi-Agent RAG research, and Systems Engineering.
Delhi Technological University, Delhi
High School Honors2021 – 2023

Delhi Police Public School

CBSE Senior Secondary (Class XII) & Secondary (Class X)

Class XII: 88.40% | Class X: 92.40%

Science & Mathematics Stream

  • ›Class XII (2023): Secured 88.40% in CBSE Board Examinations.
  • ›Class X (2021): Secured 92.40% in CBSE Board Examinations.
  • ›Awarded Academic Excellence Award for top 1% scholastic performance.
New Delhi, India

Relevant Courses & Subjects

† indicates currently ongoing course
Data Structures & AlgorithmsDatabase Management SystemsOperating SystemsComputer NetworksTheory of ComputationArtificial IntelligenceSoftware EngineeringInformation & Network SecurityCryptography & Network SecurityProbability & StatisticsLinear Algebra & ODEStochastic ProcessesAnalysis & Design of AlgorithmsNumber TheoryOptimization TechniquesMathematical Modeling
06. CONNECT

Let's Build Something Impactful

Whether you're looking to discuss Agentic AI workflows, RAG architecture, full-stack backend systems, or engineering collaborations, I'd love to connect.

+91-9310492342