PK Pratyush Pandey
+ [00]ENGINEERING PORTFOLIO

PratyushPandey

>AI & LLM SYSTEMS ENGINEER_

Specializing in AI model evaluation, data preparation, scalable RAG pipelines, and production-grade full-stack architectures from foundational database schemas to fluid reactive interfaces with Python, Next.js, React, Node.js, SQL, and LLMs.

04+
Years Dev CraftPython, TypeScript, SQL
10+
Production AppsFull-Stack & Applied AI
<38ms
P95 API LatencyFastAPI & Next.js Endpoints
99.9%
Uptime StandardFault-Tolerant Architectures
++++
SYS.COCKPIT // LIVE RUNTIME<38ms P95
END-TO-END DISTRIBUTED PIPELINE
12ms
Client TierNext.js 14 / React 18SSR + Streaming SSE • Port :3000
18ms
API & Real-Time GatewayPython FastAPI / Node.js & SocketsJWT Auth, RBAC & Socket.IO • Port :8000
38ms
AI SwarmGemini 1.5 & Agentic RAGTool-Calling Multi-Agent Swarm
3.8ms
Data TierMongoDB Atlas & Redis 7ACID Sharded + Cache Layer • Port :6379
01 CLIENT02 GATEWAY03 AI SWARM04 STORAGE
Client Tier • Next.js 14 / React 18

SSR + Streaming SSE • Port :3000 • Throughput: 3.4k req/s

12ms P95
ENVIRONMENT: PRODUCTION // MULTI-CLOUD
SECURED HTTP/2 • TLS 1.3
00 // CORE ARCHITECTURE → 01 // ABOUT ME
01 / ABOUT ME

Engineering with intent. Driven by curiosity.

Bridging the gap between beautiful user interfaces and high-scale resilient backends.

+ [PORTRAIT.01 // ARCHIVE]KL UNIVERSITY // 2026 +
Pratyush Kumar Pandey - Software Developer & AI Builder
Available for Roles
Full-Stack & AI Builder
KLEF • CGPA 8.62 • '27
Vijayawada, India• System Architecture & LLMs
ENGINEERING PROFILE & PHILOSOPHY

Hello! I'm Pratyush Kumar Pandey, a Computer Science & Engineering student at Koneru Lakshmaiah Education Foundation (KLEF) with an 8.62/10 CGPA.

I specialize in AI model evaluation, data preparation, RAG pipelines, and building practical AI-powered applications. My technical toolkit spans Python, SQL, Next.js, React.js, Node.js, and MongoDB, strengthened through an AI internship at EtharaAI and award-winning full-stack projects.

Whether evaluating LLM response quality against structured criteria, building scalable ETL data pipelines with MySQL & Power BI, or deploying real-time collaborative architectures, I enjoy using AI and modern software engineering to solve real-world problems with high reliability.

00+Verified Industry Credentials
0+Full-stack & AI projects built
0.62Academic CGPA / 10 at KLEF
0%Dedication to performance & quality
🏏 Cricket Enthusiast🏃 Marathon Runner🎧 Deep Focus Music⚡ Lifelong Learner
02 / WHAT I DO

+ [DOMAIN 01 // 01 / BUILD]SYS.01 // ACTIVE +
(01)01 / Build

Full-Stack & Systems Development

Engineering scalable web applications from foundational schemas and secure JWT authentication to high-performance REST APIs and fluid, responsive interfaces.

Next.js & Node.jsCore Stack
MongoDB & MySQLDatabase
WebSockets / Socket.IOReal-Time
Domain 01 // Core Full-Stack
PythonJavaScriptTypeScriptReact.jsNext.jsNode.jsExpress.jsMongoDBREST APIs
Core Deliverables & Pipelines
01Full-Stack Next.js & React ApplicationsFull-Stack

Interactive React.js & Next.js SPAs powered by TypeScript, Node.js/Express backends, and high-availability MongoDB databases.

02RESTful API & Microservices EngineeringBackend

Clean API design with standardized error models, rate limiting, JWT token security, role-based access control (RBAC), and Postman test suites.

03Real-Time WebSocket & Socket.IO CollaborationReal-Time

Bidirectional synchronized multi-user messaging, live audio/video channels, and real-time state synchronization.

Production FastAPI / Python Real-Time Evaluation Endpoint LIVE
@app.post("/api/v1/interviews/evaluate", response_model=EvaluationResponse)
async def evaluate_interview_stream(payload: EvaluationPayload):
    # Retrieve contextual vector embeddings & audit LLM response
    context = await vector_store.similarity_search(payload.query, top_k=5)
    evaluation = await ai_evaluator.grade(payload.audio_transcript, context)
    return {"status": "SUCCESS", "score": evaluation.accuracy, "latency_ms": 38}
+ [DOMAIN 02 // 02 / THINK]SYS.02 // ACTIVE +
(02)02 / Think

AI Engineering & LLM Systems

Building production-grade AI systems that convert Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), prompt engineering, and model evaluation into practical value.

RAG + Hugging FaceLLM Pipeline
Post-Training QAEvaluation
NumPy & PandasData Quality
Domain 02 // Applied Intelligence
PythonLLMsPrompt EngineeringRAGHugging FaceLLM APIsModel EvaluationData PreparationNumPyPandas
Core Deliverables & Pipelines
01LLMs, Prompt Engineering & RAG SystemsGenerative AI

End-to-end RAG pipelines with semantic vector search, dynamic context injection, prompt evaluation, and multi-turn voice AI synthesis.

02AI Model Evaluation & Quality AnalysisModel QA

Structured post-training workflows evaluating accuracy, relevance, instruction following, and response quality across diverse datasets.

03Data Preparation & Machine Learning FoundationsML / Data

Preprocessing, cleaning, validation, and feature preparation with Python, NumPy, Pandas, and Hugging Face pipelines.

End-to-End RAG & Model Evaluation Pipeline LIVE
01 Input StreamPrompt & voice audio tokenization→
02 Context RetrievalVector similarity top-k search→
03 LLM InferencePrompt engineering & response stream→
04 Quality EvaluationStructured rubric & benchmark scoring
+ [DOMAIN 03 // 03 / SHIP]SYS.03 // ACTIVE +
(03)03 / Ship

Data Pipelines, DevOps & Cloud

Building resilient data foundations with clean SQL schemas, containerized Docker pipelines, Jenkins CI/CD automation, and disciplined delivery.

Docker & JenkinsContainers
ETL / ELT & MySQLPipelines
Power BI DashboardsAnalytics
Domain 03 // Data & Cloud Delivery
Data PipelinesETL/ELTMySQLSQLData AnalysisPower BIDockerJenkinsGitGitHubCI/CDPostman
Core Deliverables & Pipelines
01Data Pipelines & ETL / ELT WorkflowsData Eng

Structured data processing, validation, relational modeling with MySQL, SQL data analysis, and executive dashboards in Power BI.

02Docker Containerization & CI/CD PipelinesDevOps

Automated Jenkins and GitHub Actions workflows, container build scans, environment parity, and zero-downtime deployments.

03API Auditing, Testing & Version ControlQuality

Postman test collections, API contract validation, Git/GitHub branch workflows, and secure token access controls.

Continuous Integration & Delivery Pipeline LIVE
Data Schema Validation & SQL Integrity CheckPASSED
Model Evaluation & Automated Postman Test Suite100%
Docker Container Build & Jenkins CI PipelineVERIFIED
Cloud Production Deployment & Health MonitoringLIVE
03 / SELECTED WORK

Selected Work Products I've built to solve real problems.

A selection of production systems where product thinking meets engineering craft.

03 Production Deployments Speech AI & LLM Systems Sub-50ms Real-Time Sockets 2025 — 2026
https://cogniview.ai/practice/voice-interview
Streaming Active
LIVE SESSION #0482 GEMINI 1.5 PRO STREAM • 42ms
AI Technical Prompter // Event-Driven Microservices
“Evaluate candidate response on distributed consensus, system tradeoffs, and communication depth.”
Problem Solving96%
Technical Depth94%
Voice Response95%
2026Voice AI • Placement Practice
Next.jsTypeScriptFirebaseVAPIREST APIsVercelLLMs

Cogniview: AI Voice Interview & Placement Platform

Next.js • TypeScript • Firebase • VAPI • REST APIs • Vercel

An AI-powered voice interview platform that enables users to practice interviews through interactive voice conversations and receive real-time feedback. Implemented LLM-based scoring and personalized feedback systems to evaluate communication, technical responses, and overall interview performance.

https://nexthire.ai/recruiter/pipeline
Live WebSocket Connected • 32ms48 Candidates In Review
EY
Eren YeagerSenior Full-Stack Engineer • Berlin, DE • Remote
98% MATCHVERIFIED TALENT
Recruiter Pipeline Chat:

“Your distributed systems architecture test was exceptional! Let's schedule the final round.”

WebSocket Real-Time ActivitySub-50ms Event Sync
View ↗
02.
2025AI Recruitment • WebSockets Architecture
React.jsNode.jsExpress.jsMongoDBJWTREST APIsWebSockets

NextHire: AI Job Portal & Recruitment Workspace

React.js • Node.js • Express.js • MongoDB • JWT • REST APIs • WebSockets

A scalable AI-powered job portal with secure JWT authentication, role-based access control (RBAC), real-time communication, and RESTful APIs for recruiters and job seekers. Built intelligent recruitment features including AI-based resume parsing, semantic resume-job matching, and personalized recommendations.

https://codesync.dev/workspace/collab
CodeSync Engine v2.0 • 3 Peers Live Sync: 18ms
03Peers
Python + Django + Socket.IOSynchronized Multi-Cursor CanvasSub-50ms Keystroke Sync • JWT RBAC
Active Room Collaborators:
Pratyush (Owner / Host)Alex K. (Collaborator)Elena R. (Reviewer)
✦ AI Copilot:LLM Copilot: Refactored async event handler to use Socket.IO non-blocking broadcast.
View ↗
03.
2025Collaborative Systems • LLM Code Assistance
PythonDjangoMongoDBSocket.IOJWTREST APIsLLMs

CodeSync: Collaborative Code Editor & AI Workspace

Python • Django • MongoDB • Socket.IO • JWT • Backend APIs

A scalable real-time collaborative code editor using Python and Django, featuring secure JWT authentication, synchronized multi-user code editing, role-based access control (RBAC), and RESTful APIs for seamless collaboration. Integrated LLM-powered code assistance, real-time group chat, and video calls using Socket.IO.

04 / MY JOURNEY//CHRONOLOGICAL ARCHITECTURE

Where I'velearned and grown.

00 // CHRONOLOGY

Where Ideas Met Execution

2024 — 2026 Journey

5 engineering milestones connected along an evolving path. From hackathon architectures to production AI pipelines and distributed cloud microservices.

Scroll down to travel timeline →
2026AI & Systems Engineering
CALIBO

Calibo

AI Trainee

PythonMachine LearningGenerative AILLMsRAG

Developing and testing production-grade AI solutions, enterprise LLM pipelines, and automated agent workflows to solve critical business problems with high reliability.

  • Spearheaded enterprise RAG pipeline development with vector retrieval reducing latency by 40%
  • Implemented modular tool-calling agents for automated data analysis and schema verification
  • Worked on practical software and AI use cases, building solutions for real-world business problems.
AI & Systems Engineering01 of 05
Jan 2026 – May 2026LLMs, RAG, Model Evaluation & Data Annotation
ETHARAAI

EtharaAI

AI Intern

PythonLLMsRAGModel EvaluationPrompt Engineering

Contributed to Large Language Model (LLM) post-training workflows through structured data preparation, response evaluation, and quality analysis to improve model performance and reliability.

  • Streamlined data processing & validation routines, improving consistency across evaluation datasets
  • Conducted fine-grained qualitative evaluations on LLM prompt-completion pairs across complex domains
  • Contributed to Large Language Model (LLM) post-training workflows through data preparation, response evaluation, and quality analysis to improve model performance and reliability.
LLMs, RAG, Model Evaluation & Data Annotation02 of 05
2025Smart India Hackathon, HackVega & Tata Imagination Finalist
HACKATHONS

National Hackathons // NextHire

Full-Stack & AI Builder

React.jsNode.jsExpress.jsMongoDBJWT

Built NextHire — an end-to-end intelligent recruitment platform featuring AI resume parsing, live recruiter-candidate chat, automated candidate screening, and recruiter dashboards.

  • Selected as finalist and competitive participant across Smart India Hackathon, HackVega, and Tata Imagination Challenge
  • Engineered real-time chat supporting instant presence and typing indicators with 0 dropped frames
  • Engineered full-stack architecture using React.js and Node.js with secure token-based authentication and role-based access control (RBAC).
Smart India Hackathon, HackVega & Tata Imagination Finalist03 of 05
2024 – 2025Real-Time Distributed Code Architecture
SYSTEMS

Collaborative CodeSync Systems

Full-Stack & Systems Developer

PythonDjangoMongoDBSocket.IOJWT

Engineered CodeSync — a scalable real-time collaborative code editor with synchronized multi-user code editing, secure JWT authentication, role-based access control (RBAC), and RESTful APIs.

  • Achieved sub-50ms keystroke synchronization between distributed peers using Socket.IO broadcasting
  • Implemented conflict-free replicated data structure handling concurrent typing smoothly
  • Developed a scalable real-time collaborative code editor using Python and Django featuring secure JWT authentication and role-based access control.
Real-Time Distributed Code Architecture04 of 05
2023 – Expected Aug 2027Academic Honors // CGPA: 8.62 / 10
KLEF

Koneru Lakshmaiah Education Foundation

B.Tech in Computer Science & Engineering

AI/MLDBMSMERN Stack DevelopmentSoftware EngineeringPython

Pursuing Bachelor of Technology in Computer Science and Engineering with an exemplary 8.62/10 CGPA. Completed rigorous coursework in AI/ML, DBMS, MERN Stack Development, and Software Engineering.

  • Ranked in top academic percentile of the Computer Science & Engineering department
  • Active participant in collegiate coding contests and technical innovation challenges
  • Maintained an outstanding 8.62/10 CGPA across all semesters of engineering studies.
Academic Honors05 of 05
05 // UP NEXT

Verified Credentials

Certifications & Standards

Deep-dive into verified industry credentials spanning MongoDB developer certifications, Oracle Cloud Infrastructure, Cambridge English, and GitHub Foundations.

View Credentials
Intro: Chronology// 2024 — 2026 Journey
00 // Start0% Explored
04 / MY JOURNEY

Where I'velearned and grown.

00
00 // CHRONOLOGY

Where Ideas Met Execution

2024 — 2026 Journey

5 engineering milestones connected along an evolving path. From hackathon architectures to production AI pipelines and distributed cloud microservices.

01
2026
CALIBO

Calibo

AI Trainee

PythonMachine LearningGenerative AILLMsRAG

Developing and testing production-grade AI solutions, enterprise LLM pipelines, and automated agent workflows to solve critical business problems with high reliability.

  • Spearheaded enterprise RAG pipeline development with vector retrieval reducing latency by 40%
  • Implemented modular tool-calling agents for automated data analysis and schema verification
  • Worked on practical software and AI use cases, building solutions for real-world business problems.
AI & Systems Engineering01 of 05
02
Jan 2026 – May 2026
ETHARAAI

EtharaAI

AI Intern

PythonLLMsRAGModel EvaluationPrompt Engineering

Contributed to Large Language Model (LLM) post-training workflows through structured data preparation, response evaluation, and quality analysis to improve model performance and reliability.

  • Streamlined data processing & validation routines, improving consistency across evaluation datasets
  • Conducted fine-grained qualitative evaluations on LLM prompt-completion pairs across complex domains
  • Contributed to Large Language Model (LLM) post-training workflows through data preparation, response evaluation, and quality analysis to improve model performance and reliability.
LLMs, RAG, Model Evaluation & Data Annotation02 of 05
03
2025
HACKATHONS

National Hackathons // NextHire

Full-Stack & AI Builder

React.jsNode.jsExpress.jsMongoDBJWT

Built NextHire — an end-to-end intelligent recruitment platform featuring AI resume parsing, live recruiter-candidate chat, automated candidate screening, and recruiter dashboards.

  • Selected as finalist and competitive participant across Smart India Hackathon, HackVega, and Tata Imagination Challenge
  • Engineered real-time chat supporting instant presence and typing indicators with 0 dropped frames
  • Engineered full-stack architecture using React.js and Node.js with secure token-based authentication and role-based access control (RBAC).
Smart India Hackathon, HackVega & Tata Imagination Finalist03 of 05
04
2024 – 2025
SYSTEMS

Collaborative CodeSync Systems

Full-Stack & Systems Developer

PythonDjangoMongoDBSocket.IOJWT

Engineered CodeSync — a scalable real-time collaborative code editor with synchronized multi-user code editing, secure JWT authentication, role-based access control (RBAC), and RESTful APIs.

  • Achieved sub-50ms keystroke synchronization between distributed peers using Socket.IO broadcasting
  • Implemented conflict-free replicated data structure handling concurrent typing smoothly
  • Developed a scalable real-time collaborative code editor using Python and Django featuring secure JWT authentication and role-based access control.
Real-Time Distributed Code Architecture04 of 05
05
2023 – Expected Aug 2027
KLEF

Koneru Lakshmaiah Education Foundation

B.Tech in Computer Science & Engineering

AI/MLDBMSMERN Stack DevelopmentSoftware EngineeringPython

Pursuing Bachelor of Technology in Computer Science and Engineering with an exemplary 8.62/10 CGPA. Completed rigorous coursework in AI/ML, DBMS, MERN Stack Development, and Software Engineering.

  • Ranked in top academic percentile of the Computer Science & Engineering department
  • Active participant in collegiate coding contests and technical innovation challenges
  • Maintained an outstanding 8.62/10 CGPA across all semesters of engineering studies.
Academic Honors05 of 05
06
05 // UP NEXT

Verified Credentials

Certifications & Standards

Deep-dive into verified industry credentials spanning MongoDB developer certifications, Oracle Cloud Infrastructure, Cambridge English, and GitHub Foundations.

View Credentials
04 // JOURNEY ARCHITECTURE → 05 // CREDENTIALS
05 / CREDENTIALS

Always learning, always leveling.

Industry-recognized certifications and verified credentials.

06 / GET IN TOUCH

Have an idea? Let's make it real.

Whether it's a new opportunity, an engineering collaboration, or a chat about AI and full-stack systems — I'd love to connect.

Start a conversation
01

Let's build something
useful, thoughtful, and memorable.

Email Addresspratyush.me.ai@gmail.comDirect inbox response
Direct Phone+91-7562920811
Based InVijayawada, IndiaAndhra Pradesh, India
Local Time--:-- ISTAsia/Kolkata (UTC +05:30)
Find Me Online
02

Tell me about your project

I usually reply within 24 hours.

0/3000 (min 10)
Every good collaboration starts with a conversation.