MODULAR BIOS v6.00PG
Energy Star Ally
Press DEL to enter SETUP
15/07/2026
Parth Deshpande
System Administrator // AI Engineer
Secure Workstation Locked
about_parth.txt
WHO_AM_I.TXT
Parth Deshpande
B.Tech CSE (AI & ML) @ VIT Bhopal
Systems & Machine Learning Engineer
terminal — zsh
task_manager
Running ProcessesTasks: 4 active
custom-hnsw-coreRunning
vitamin-api-serverActive
netlyze-packet-inferenceActive
ollama-llama3.2-streamRunning
HOME PROJECTS CAREER INTEL REPORT

PARTH
DESHPANDE
SYSTEMS & ML

B.Tech CSE (Artificial Intelligence & Machine Learning) · Vector Databases · Distributed Backends · Real-time Systems
STATUS: COMPILING_CORE
INITIATE_CONTACT()
◈ LIVE_DATA
SYSTEM_METRICS
MEASURED. REAL. VERIFIABLE.
0
DSA Problems Solved
Mastered across LeetCode, Codeforces, & CodeChef.
Top 2.2%
Amazon ML School
Shortlisted from 1.34 Lakh+ registrations nationwide.
1409
Max Codeforces Rating
Attained verified Pupil Rank status engine.
0
Active Users Optimized
VITAMIN EdTech scale baseline contract capability.
CAREER_LOG ACCESS_LEVEL: COMPILING
[Aug 2025 – Mar 2026] FOUNDING
Co-Founder & Backend Lead
@ VITAMIN EdTech📍 Bhopal, India
  • Co-founded and engineered the FastAPI backend infra supporting a campus microservices node scaling over 20,000+ students.
  • Designed SQLite-based offline sync paradigms alongside server reconciliation strategies via real-time Socket.io threads.
  • Built independent dashboard, scheduling, and chat service limits integrating conversational LLM peer discovery.
MicroservicesFastAPISocket.ioMongoDBArchitecture
[Dec 2024 – Feb 2025] INTERNSHIP
A.I. Developer Intern
@ AlgoFed📍 Remote
  • Engineered low-level machine learning pipelines and custom automated execution models.
  • Built high-throughput processing scripts parsing, tokenizing, and evaluating unstructured model context.
  • Collaborated within the core internship matrix handling real-world deployment limitations.
AI DeveloperPythonData ParsingAutomation
[Sep 2023 – Oct 2023] INTERNSHIP
C++ Developer Intern
@ Code Casa Pvt. Ltd.📍 Remote
  • Developed core system utilities and multi-threaded algorithm patterns utilizing standard C++ environments.
  • Implemented specialized object visualization layouts and performance verification matrices.
  • Optimized low-level computational execution boundaries reducing raw overhead metrics.
C++ EngineeringData VisualizationAlgorithms
// EDUCATION_LOG
[2023 – 2027] DEGREE
B.Tech – Computer Science & Engineering
Specialization: Artificial Intelligence & Machine Learning
@ Vellore Institute of Technology (VIT) Bhopal
AI/MLSystemsB.TechCGPA: 9.00
CERT
VALID
INTEL REPORT // TRACKED SYSTEMS
INTEL_REPORT
amazon_mlss_certificate.png
Amazon ML Summer School 2026 Certificate
// CERTIFICATE OF COMPLETION — AMAZON ML SUMMER SCHOOL '26
mlss_swags.jpeg
Amazon ML Summer School swags received
// SWAGS RECEIVED FROM AMAZON
mlss_learnings.log
Amazon ML Summer School is Amazon's flagship, industry-led machine learning program for students across India. I was selected among 1.34 Lakh+ registrations (top 2.2%) and completed the full 9-module curriculum — from classical supervised learning through deep learning, generative AI & LLMs, agentic systems, and causal inference. Here's what I learned:
01Supervised Learning
Regression and classification fundamentals, loss functions, regularization, bias–variance trade-off, tree-based models and ensembles, and rigorous model evaluation.
02Deep Neural Networks
Backpropagation, activation and optimization strategies, convolutional architectures, and practical training techniques for stable, well-generalizing networks.
03Dimensionality Reduction
Handling high-dimensional data with PCA and related methods, feature extraction, and learning compact representations for visualization and efficiency.
04Unsupervised Learning
Clustering, density estimation, and anomaly detection — discovering structure in unlabeled data and judging the quality of what is found.
05Sequential Learning
Modeling time series, text, and other ordered data using recurrent architectures and attention-based approaches that capture long-range dependencies.
06Reinforcement Learning
Markov decision processes, value- and policy-based methods, and the exploration vs. exploitation trade-off in learning from interaction and rewards.
07Generative AI & LLMs
How generative models and transformer-based large language models are pre-trained, adapted, and prompted — and how they are applied in real products.
08Agentic AIBONUS
Building LLM-driven agents that plan, call tools, and execute multi-step workflows autonomously, along with the design patterns that make them reliable.
09Causal Inference
Moving beyond correlation — causal graphs, confounders, and treatment-effect estimation to reason about the real impact of decisions and interventions.
// RECENT_ACHIEVEMENT
adobe_student_insider.png
Adobe Student Insider 2026-2027 badge
// OFFICIAL ADOBE STUDENT INSIDER 2026–2027 BADGE
adobe_insider_briefing.log
Selected as one of 100 Adobe Student Insiders in India for the 2026–2027 academic year — chosen from nearly 19,000 student applicants. Adobe expanded this year's India cohort to 100 students in response to the overwhelming interest, and my application stood out among them.
Program
Adobe Student Insider — India
Cohort
2026–2027 school year
Selectivity
100 selected from ~19,000 applicants
Kick-off
Adobe Education Summit · Noida Office · Oct 6
  • Selected on the strength of VITamin — the EdTech platform now used by 20,000+ students at VIT Bhopal, where I architected the backend and worked across the React Native and web front ends.
  • Recognized for designing around a real constraint — patchy campus Wi-Fi — by building offline-first persistence so students can keep using the app when the network drops: designing for the bad case, not just the demo case.
  • Over the year: connect and co-create with Adobe teams, share product feedback and ideas, and collaborate with fellow Insiders from across the country.
  • Begins with the Adobe Education Summit at Adobe's Noida office, followed by year-long community and co-creation opportunities.
// OPERATIONAL_ACHIEVEMENTS
FinalistNATIONAL
HackTU 6.0
Top engineering squad out of 800+ competing clusters (3,500+ participants). Built an AI-Powered Lending Engine.
Thapar Institute of Eng. & Tech
PupilALGO
Codeforces Ranked
Max metric ranking 1409 verified. High competence tracking execution.
Codeforces Index
5-Star SolverCOMPUTE
HackerRank Command
5-Star Badge earner across C++ and Python problem domains.
HackerRank Platform
Silver BadgeALGO
CodeChef Problem Solver
Advanced algorithmic runtime debugging and structural validation badge.
CodeChef
// INTEL_LITERATURE
TECHNICAL_PUBLICATION
Python for Machine Learning: The Complete Roadmap Nobody Told You About
A comprehensive, engineering-first strategic roadmap parsing optimization paths, algorithmic matrix libraries, and deployment models often obfuscated for standard developers. Published on Hashnode.
READ_PUBLICATION_LOG_↗
// VERIFIED_CERTIFICATIONS (Interactive Slide-down Trigger Nodes)
ICTRD Data Science Certificate
Cognizant Introduction to Agentic AI Certificate
Cognizant Prompt Engineering Foundations Certificate
Cognizant Fundamentals of Generative AI Certificate
Oracle Agentic AI Certificate
OCI 2025 AI Associate
Google Kaggle AI Agents Certification
University of Michigan ML Certificate
NPTEL IoT Certificate
NPTEL Cloud Computing Certificate
AlgoUniversity Advanced Training Core
AWS Technical Essentials Validation
Bosscoder Advanced SQL Certificate
CORE ARSENAL
THE SECRET SAUCE
TARGET ARCHIVES // ENCRYPTED CORE
MISSION
ARCHIVES
>ACCESSING REPOSITORY FILES...

Declassified operational blueprints engineered by Parth Deshpande. Deep tech infrastructure, AI, and raw systems code.
SYSTEM RECONCILED DECRYPTION STATUS
EVIDENCE_#001
Autonomous Production Choke Controller
▶ [ DECRYPT_VISUALS ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.08
OPERATION:
Autonomous Production Choke Controller — MPC Digital Twin
Closed-loop autonomous Model Predictive Control system and digital twin for oil well choke valves. Uses SLSQP optimization and dynamic system identification (R² > 0.993) to maximize production while enforcing WHP, FLP, and BHP safety envelopes and actuator ramp limits.
PythonMPCSLSQPSystem IDStreamlit
EVIDENCE_#002
Vector Database
▶ [ DECRYPT_VISUALS ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.07
OPERATION:
Vector-Database — High-Dimensional Search Engine
Engineered a custom vector database completely from scratch in C++, implementing HNSW, KD-Tree, and multi-metric search pipelines (Cosine, Euclidean, Manhattan) alongside local LLM RAG pipelines.
C++HNSWKD-TreeOllamaRAG
EVIDENCE_#003
VITAMIN App
▶ [ VIEW_SOURCE ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.03
OPERATION:
VITAMIN — Microservices Campus Core
Architected the FastAPI backend and MongoDB schema for an enterprise ed-tech environment scaling across 20,000+ nodes. Integrated an intelligent LLM peer-matching core.
FastAPIMongoDBReact-NativeGemini APISocket.io
EVIDENCE_#004
Netlyze System
▶ [ VIEW_SOURCE ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2025.03
OPERATION:
Netlyze — Traffic Pipeline Optimizer
Built a high-performance predictive packet processing engine bridging C++ packet inspection with isolated Python ML diagnostic runtimes for predictive traffic evaluation.
C++PythonScikit-LearnSQLWireshark
EVIDENCE_#005
Plagiarism Detector
▶ [ VIEW_SOURCE ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.03
OPERATION:
Semantic Code Similarity — Plagiarism Core
Designed an NLP-powered system analyzing source files at the semantic structural tier rather than structural text similarity, tracking complex rename masking strategies.
PythonWord2VecGloVeNLPStreamlit
EVIDENCE_#006
HCI System
▶ [ VIEW_SOURCE ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.01
OPERATION:
Subway Surfers Virtual Play — Gesture HCI
Engineered a low-latency Human-Computer Interaction pipeline translating structural body physics landmark data directly to interactive system controls.
PythonOpenCVMediaPipePyAutoGUI
EVIDENCE_#007
Netninja Web Engine
▶ [ CONNECT_TO_ENGINE ]
CLICK_TO_ACCESS
CONFIDENTIAL// DATE: 2026.05
OPERATION:
Netninja — Admin Security Orchestration Console
Developed a unified, high-integrity dashboard system structured with client-side reactive components for specialized node administration and administrative analytics oversight.
React.jsViteTailwindCSSFirebaseSecurity Console
SIGNAL_INTERCEPT // PARTH_NODE
Global Intel
>DECRYPTING SYSTEMS RUNTIME METRICS...
TARGET_NODE_ID
Parth-bot-crypto26
LEETCODE
deshp_parthALGORITHMIC_MATRIX
CODEFORCES
Ash_26PUPIL_RANK
CODECHEF
parthdeshpandeCOMPETITIVE_ENGINE
CODOLIO
Ash_26UNIFIED_PROFILE
COMPUTE_FREQUENCY: SYNCEDDATA_INTEGRITY: SECURE
ESTABLISH
CONNECTION
Transmit architecture metrics or orchestration parameters securely.
>_ COPY_ENCRYPTION_KEY (EMAIL)
deshpandeparth793@gmail.com
TRANSMISSION_SOURCE: INDIA (IST) STATUS: ACCEPTING_COMPUTATIONAL_CHALLENGES