Alosh Denny

• Training, finetuning and instructing agentic AI systems for cybersecurity management
• Developing autonomous AI agents capable of threat detection, analysis and response
• Implementing reinforcement learning techniques for adaptive cybersecurity protocols

• Led the development of multimodal language models specialized for Malayalam-language EdTech ap
plications.
• Automated mentor and student call QA, video search, and reasoning in Adapt AI, the in-house AI Chat.
• Designed and implemented automated pipelines on ChromaDB and Qdrant for vector database manage
ment and retrieval-augmented generation (RAG) integration.
• Utilized Modals serverless cloud infrastructure to efficiently deliver AI-generated content to students.
• Spearheaded the implementation of Adapt AI, leading to:
- 21% increase in Daily Active Users
- 23% increase in Weekly Active Users
- Nearly 50% growth in Monthly Active Users within one week of launch
• Contributed to the apps marketplace success, resulting in:
- Improving app ratings from 4.6 to 4.8
- Ranking as the 1 educational app on Play Store in Kerala
- Ranking among the top 50 apps on the App Store

• Built an autonomous Martian Rover for the European Rover Challenge
• Martian Geological Research via JMARS
• Autonomous Drone built upon MATLAB, Raspberry Pi and CARLA Sim

• Modelled and forecasted import/export traffic for Cochin Port up to 2030
• Worked closely with numerous statistical and Deep Learning algorithms
• Utilized ensemble architectures such as LSTM, RNN, Prophet, XGBoost, CatBoost, and NGBoost for pre
dictive modeling
• Trained models using grid search to optimize hyperparameters for better accuracy
• Devised a novel combinatorial architecture combining Deep Learning techniques and LSTM for improved
traffic forecasting