DARABODLA
3rd year AI student at GIKI. Co founder and CEO of CPC. Building full stack AI systems from training to production, one commit at a time.
WHO IS
DARA BODLA?
AI student by day, founder by night. Studying at GIKI (BSc in Artificial Intelligence) and running CPC, a WhatsApp automation startup serving restaurants, clinics, and travel agencies. I ship end to end AI systems: computer vision, NLP, MLOps, and full stack apps. I interned at Corbis Soft building YOLOv8 pipelines and DQN chess agents, and have shipped 10+ production grade projects.
"It's not who I am underneath… but what I build that defines me."
Shipped 3 end to end AI prototypes in 10 weeks: a YOLOv8 floor-plan dimension reader (mAP 0.812), a Deep Q-Network chess agent (win rate 0→0.437), and a role based ERP system with PostgreSQL.
Built WhatsApp automation platform serving 3+ business verticals. Sub 2s latency, live Meta catalogue integration, custom onboarding portal, billing system.
Organised Hack Nation × MIT hackathon. Built sponsorship decks, interactive HTML presentations, and "The AI Scientist" project for the World Bank Youth Summit track.
SKILLS
& TECH
Hover any tag
for context.
NOTABLE WORK
Silent speech recognition using 3D-CNN, MS-TCN and BiLSTM on the LRW dataset. 11 modules and 2900 lines of production code trained on T4 GPU with curriculum learning.
Full MLOps pipeline. XGBoost ensemble, FastAPI, Prefect, Docker, GitHub Actions CI/CD.
Faster R-CNN MobileNetV3. Real time webcam inference, 87.31% mAP.
3 complete AI prototypes shipped in 10 weeks. From zero to working demo for each. YOLOv8 floor-plan dimension reader, Deep Q-Network chess agent, and a role-based ERP system.
WhatsApp automation SaaS co-founded in 2025. Meta catalogue, billing, sub-2s latency across restaurants, clinics and travel agencies.
Serverless AWS pipeline: S3 triggers Lambda, Transcribe, Translate, Polly. English lectures become Urdu audio. Auto-scales to zero.
LSTM closed-loop QoS controller on 23-node backbone. 90.6% fewer congestion events (468 to 44). Sub-12ms RTT. Autonomous retraining every 500 samples with drift detection.
473-line CloudFormation. 20+ resources. Deploy time 45 min to 5 min.
Hypothesis-to-experiment pipeline using Semantic Scholar API for live paper grounding. Built for the Hack Nation x World Bank Youth Summit. Exports a full structured experiment plan in PDF or JSON.