Gagan
Ahlawat
Software Development Engineer & Systems Researcher · SDE Intern @ Inforida Technologies. Specialized in high-concurrency backend architecture, RF wireless signal processing (CrossNet-MIMO), and open source MCP integration layers.
About & Core Ethos
"I don't ship what I don't understand."
Dual Degree (B.Tech IT + MBA) student at ABV-IIITM Gwalior. SDE Intern at Inforida Technologies and Research Intern at Department of Management Studies. Specialized in high-concurrency backend architecture, MIMO channel estimation neural models (CrossNet-MIMO), and open source MCP integration layer.
- Backend RBAC Tiers4 Levels
- LeetCode Solved300+
- LeetCode Rating~1496
- SAC TenureSecretary
Engineering & Leadership Sequence
Featured Systems & Projects
CodeLink
Real-time Mobile Code Review & VS Code Extension Relay
Real-time mobile code review relay and VS Code pairing platform with token-based handshakes and zero message loss under concurrent load.
- ›Designed a Socket.IO relay for real-time mobile code review, achieving sub-100ms sync latency.
- ›Implemented token-based handshake and QR-code pairing, blocking 0-tolerance unauthorized access attempts.
CloudOptRL
4D MDP Reinforcement Learning Cloud Allocation
Cloud resource allocation modeled as a 4D Markov Decision Process with property-based test validation.
- ›Modeled cloud resource allocation as a 4D Markov Decision Process with 3 discrete actions.
- ›Formulated a normalized reward function balancing a 40–70% target utilization band against cost.
OllamaTUI
Arch-Native Terminal UI for Local LLM Orchestration
Developer-first terminal interface for local LLMs powered by Go, Bubble Tea, and non-blocking token streaming.
- ›Shipped a developer-first TUI for local LLM inference, supporting 5+ hosted models via Ollama.
- ›Cut perceived response latency by ~60% with non-blocking token streaming via Go goroutines.
CrossNet-MIMO
Multi-Task MIMO Channel Estimation Neural Network
Packed multi-task feature extraction architecture for 64-antenna MIMO channel estimation and localization.
- ›Developed CrossNet-MIMO, packing multi-task feature extraction into a 4.8M-parameter model.
- ›Trained MIMO channel estimation models on 10,000+ samples, driving NMSE down to 0.47–0.52.
// Additional Engineering Projects
Research & Open Source Contributions
Research & Technical Papers
CrossNet-MIMO Channel Estimation & Localization Model
Deep learning multi-task feature extraction architecture for MIMO wireless channel estimation.
- Developed CrossNet-MIMO, packing multi-task feature extraction into a 4.8M-parameter model.
- Trained MIMO channel estimation models on 10,000+ samples, driving NMSE down to 0.47–0.52.
- Optimized the localization pipeline, cutting estimation error by 3–5× over baselines.
4D MDP Cloud Resource Allocation Engine (CloudOptRL)
Reinforcement learning agent optimizing cloud utilization with property-based test validation.
- Modeled cloud resource allocation as a 4D Markov Decision Process with 3 discrete actions.
- Formulated a normalized reward function balancing a 40–70% target utilization band against cost.
- Constructed an EpisodeGrader with 30/30/40% weighted scoring, validated via property-based tests.
Open Source Ecosystem
Resolved setup.py dependency bug blocking pip installs (#360 fix) and migrated dependency management to pyproject.toml for Python 3.12 support across 3 merged PRs.
Built global repository search feature extending discovery to any public repo, delivered the MCP integration layer, and fixed 3 UI bugs across a 500+-commit codebase.
Active contributor to Linux systems tools, security benchmarking scripts, and Python optimization libraries.