I build AI-powered software end to end: agents that plan, pipelines that render, and full-stack products people can actually use.
class Developer:
def __init__(self):
self.name = "Gokul A"
self.education = "B.Tech Computer Science — Crescent College [Fresher — 2026 batch, ready to join immediately ]"
self.internships = [
"AI/ML Developer Intern @ Infosys Limited (Aug–Oct 2025)",
"Full Stack Developer Intern @ Zidio Development (Jul–Sep 2024)"
]
self.flagship = "ShortForge — AI Video Generation Factory [Active]"
self.recent_projects = ["HireHawk — Autonomous AI Job Application Copilot",
"Sentixcare — Mood-driven personalized recommendation system"]
self.currently = "Building ShortForge — agentic orchestration + distributed GPU rendering"
self.exploring = ["Temporal workflows", "ROCm / AMD AI compute", "Agent verification & evals"]
self.open_to = "Entry-level AI/ML · Software Engineering · Full Stack roles (2026)"
self.locations = "Chennai, Tamil Nadu · open to remote / relocation"
self.publications = ["Mood-driven multimodal recommendation systems (ICRIT '26)"]
self.achievements = ["Smart India Hackathon (SIH) 2025 Finalist & Internal Winner"]
self.core_stack = ["PyTorch", "LangGraph", "LangChain", "MCP",
"TypeScript", "React.js", "FastAPI", "Docker", "AWS"]
self.certifications = ["AWS AI Practitioner Challenge — Udacity (Jun 2026)",
"CLLMSP — LLM Security (95%)"]
Live numbers from my GitHub account — updated automatically
🟩 Contribution Graph
Every project below has a live demo and source code · Start with the flagship, then browse the rest
| Project | What it is | Domain | Status | Links |
|---|---|---|---|---|
| ⚙️ ShortForge | AI short-form video generation platform | Agentic AI · Distributed GPU | 🟢 Active | Live · Code |
| 🦅 HireHawk | Autonomous job-application copilot | Multi-agent · Chrome extension | 🟢 Live | Live · Code |
| 🧠 Sentixcare | Mood-driven recommendation system | Multimodal AI · Research | 📄 Published | Live · Code · Paper |
| 🏆 Planetopia | Gamified environmental education | MERN · SIH 2025 Finalist | 🏅 Finalist | Live · Demo · PPT |
| 🌱 Carbon Ingest | Carbon emissions data ingestion & audit | Django · React · Data pipelines | 🟢 Live | Live · Code |
|
A video "factory" where AI agents plan, GPU workers render, and verification decides what ships. |
|
|
🎯 What it does
|
🛡️ How it stays reliable
|
|
|
|
|
Mood-driven personalized recommendation system What it does
|
Autonomous AI job-application copilot What it does
|
|
SIH 2025 Finalist — gamified environmental education What it does
|
Enterprise carbon data ingestion & audit system What it does
|
A simplified view of the closed-loop idea behind the platform:
flowchart LR
A[🎬 Brief / Prompt] --> B[🧠 Ascalon<br/>Cognitive Decisioning]
B --> C[🤖 Agent Orchestration<br/>MCP Tools]
C --> D[🔀 Provider Routing]
D --> E[🖥️ Distributed GPU Rendering<br/>Leases + Fencing]
E --> F{✅ Verification Gates}
F -- pass --> G[🚀 Release]
F -- fail --> B
G -. feedback .-> B
Typed contracts connect every stage; failed verification loops back into decisioning instead of shipping bad output.
- Built a CNN + NLP hybrid for real-time mood-based music playback using OpenCV + PyTorch
- Optimized inference pipelines for high-speed webcam detection; deployed backends via Flask and Streamlit
- Owned the full AI lifecycle — data ingestion → model training → production deployment — in an Agile team
- Built a MERN-stack Job Hunt platform with RESTful APIs supporting full CRUD
- Reduced frontend latency by 35% and kept API response times under 500ms
- Delivered all milestones 2 weeks ahead of schedule in an Agile/Scrum setup
| Principle | What it looks like in my work |
|---|---|
| 🧾 Clear contracts | Agents, tools and services talk through explicit schemas |
| ✅ Verify before shipping | Outputs are checked before release, not after users complain |
| 🙋 Human in the loop | Autonomy where it's safe, approval gates where it matters (HireHawk) |
| 🛡️ Security-aware LLM work | CLLMSP-certified (95%) — prompt-injection and tool-abuse aware |
| 🚢 Ship, then iterate | Live deployments for my projects, not just repos |
Deep Learning & Computer Vision
NLP & LLM Orchestration
Open-Source LLMs
Serving & APIs
Agent & Compute Infrastructure
Languages
Frontend
Backend
Databases
DevOps & Cloud
| Year | Title | Venue | Link |
|---|---|---|---|
| 2026 | Mood-driven multimodal recommendation systems | ICRIT '26 — Conference Paper |  |
- 🥇 SIH 2025 Finalist — Planetopia (EcoSpark) | Won internal college round → National level
- 🏅 Winner — Talent Hunt Tech Quiz, Crescent College of Engineering
- 🤖 Cobotics Club Executive - Executive Lead, under the I-Hub Foundation for Cobotics (IHFC), the Technology Innovation Hub of IIT Delhi.
- 👥 Senior Sub-Committee Member (IT) — Crescent Club of Finance; led Alumni website development
- 📚 Research Publication — AI/ML domain (ICRIT '26)
 AWS AI Practitioner Challenge Udacity · Jun 2026 |
 Certified LLM Security Professional RedTeams · Score: 95% |
 Deep Learning for Developers Infosys Springboard |
 Machine Learning Foundations Infosys Springboard |
 System Design GeeksforGeeks |
 Published Researcher — ICRIT '26 Mood-driven multimodal AI |
Hiring for an entry-level AI/ML, software engineering or full-stack role? I learn fast, ship working demos, and care about reliability.
"Building production AI with LLMs, agents, multimodal systems, distributed compute, and verification."