🔨 Currently building: IoT Agri Thesis & Internship Applications

Lokesh
Kumar

Software Engineering Researcher & Lead Systems Architect | Arid Agritech, Low-Resource NLP & Socio-Technical AI Systems | Sindh Agriculture University (SAU), Tando Jam

Software Engineering Researcher & Lead Systems Architect at Sindh Agriculture University (SAU), Tando Jam. Specializing in offline-first edge IoT telemetry for precision arid agritech (Agri-IDAS), low-resource Indo-Aryan vernacular NLP (Dhatki), and empirical socio-technical AI policy frameworks across the Global South.

4
Flagship Systems
12 Featured · 40+ Repos
7+
Certifications
3.00
CGPA
IEEE
Student Member # 99402233 ↗
Lokesh Kumar — Software Engineer
Open to Internships & Roles
🔬 4 Flagship Systems · 4 DOIs

Tech Stack & Tools

🐍 Python 🗄️ SQL ⚡ JavaScript ES6+ 🌐 HTML5 / CSS3 🐼 Pandas 🔢 NumPy 📡 NetworkX 🚀 Streamlit 🗺️ Folium / Leaflet 🤖 NLP / RAG 🌀 WebSockets 🐙 Git / GitHub 📊 MS Excel 🌍 WordPress ☁️ GCP / Streamlit Cloud ▲ Vercel 🔵 Netlify 📐 Figma 🐍 Python 🗄️ SQL ⚡ JavaScript ES6+ 🌐 HTML5 / CSS3 🐼 Pandas 🔢 NumPy 📡 NetworkX 🚀 Streamlit 🗺️ Folium / Leaflet 🤖 NLP / RAG 🌀 WebSockets 🐙 Git / GitHub 📊 MS Excel 🌍 WordPress ☁️ GCP / Streamlit Cloud ▲ Vercel 🔵 Netlify 📐 Figma

Academic Research & Peer-Citable Publications

CERN Zenodo Open Research Archive

Permanent, peer-citable scientific manuscripts registered with DataCite international DOIs under the CERN Zenodo Open Research Archive. Read manuscripts directly without leaving the portfolio or inspect reproducible datasets.

📄 PREPRINT • CERN ZENODO (2026) Vernacular NLP

Benchmarking Intent Classification in Extremely Low-Resource Indo-Aryan Dialects: An Empirical Study on Dhatki Agricultural Telemetry

Lokesh Kumar (Sole Author & Lead Researcher | SAU Tando Jam)

Introduces the first 1,000-utterance agricultural intent corpus for the Dhatki dialect. Benchmarks custom vernacular BPE tokenization against transformers, achieving 89.4% Micro-F1 with 14.2 ms edge CPU latency.

📄 TECHNICAL REPORT • CERN ZENODO (2026) Edge IoT & Agritech

Resilient Edge-Computing Architectures for Precision Agritech Under Extreme Arid Conditions: The Thar Desert Telemetry Blueprint

Lokesh Kumar (Sole Author & Lead Systems Architect | SAU Tando Jam)

Documents the offline-first edge IoT telemetry framework of Agri-IDAS. Integrates sub-GHz LoRa and solar-harvested gateways with SQLite ring buffers to deliver 99.4% telemetry integrity across 12-hour blackouts and 38.6% water savings in 48°C Thar desert conditions.

📄 WHITEPAPER • CERN ZENODO (2026) Labor Economics & AI

The Broken Ladder: Generative AI, Digital Feudalism, and the Structural Collapse of Entry-Level Tech Employment for Global South Graduates

Lokesh Kumar (Sole Author | Dept. of Software Engineering, SAU Tando Jam)

An empirical investigation into how generative AI coding assistants disrupt traditional offshore IT labor arbitrage in developing economies, analyzing graduate hiring telemetry and formulating a 4-pillar sovereign AI policy framework.

📄 POLICY BRIEF • CERN ZENODO (2026) Policy & Cybercrime

Do States Eat Their Own? Predatory Debt, Digital Blackmail & Thar's Suicide Crisis: A Socio-Legal & Policy Intervention Brief

Shewa Ram Suthar (Program Manager, SPO Sindh) & Lokesh Kumar (SAU Tando Jam)

Analyzes predatory microcredit algorithms, cyber extortion by unlicensed mobile loan apps, and the systemic crisis in rural Sindh. Formulates a tripartite mitigation strategy spanning FIA cybercrime enforcement and grassroots psychological support.

⬇ Resume