Tactical FPS Brain
I like problems where timing, information, and decisions all matter.
I build applied ML systems, interpretability experiments, and rough ideas that can survive outside a notebook. Currently studying AI/ML at SRMIST with a 9.35 CGPA.

$ npm run future
compiling ideas...
AI/ML student, model tinkerer, systems thinker, and professional over-optimizer of ideas.
I’m an AI/ML engineering student interested in turning research-heavy AI ideas into runnable systems. I like models that can be tested, inspected, and explained without losing the fun of building them.
$ whoami
AI/ML student building useful intelligent systems
$ cat focus.txt
transformers | interpretable AI | autonomous systems | agents
$ status --today
testing ideas, tuning baselines, lifting heavy things
I like problems where timing, information, and decisions all matter.
Hidden structure, messy clues, and technical rabbit holes are my comfort zone.
Progressive overload applies to code, models, and training logs.
I’m interested in tools that can plan, act, and explain themselves.
SensaNet, robotics-adjacent AI, and industrial intelligence keep pulling me in.
A practical stack for ML experiments, model building, data work, and deployment.
Applied AI systems with enough detail to inspect, discuss, and improve.
2025
Voice-based cognitive decline screening using acoustic features, Whisper transcription, NLP biomarkers, and anomaly scoring.
2025
Crop-support AI prototype focused on disease signals, field context, and decision-friendly recommendations.
2025
Interpretable ML workflow for spotting student risk early through academic, attendance, and engagement signals.
2025
Two-stage skin lesion pipeline combining ResUNet segmentation with EfficientNetB0 classification on ISIC data.
2025
Desktop assistant interface built with JavaFX, OpenAI models, and a MySQL memory layer for structured conversations.
Loose, technical, and exploratory work that shapes the bigger builds.
lab/status
This is where ideas are allowed to be rough: sensing networks, model probes, technical explanations, and Kaggle experiments that sharpen the real builds.
Sensor-network AI concept for perception, anomaly detection, and autonomous feedback loops.
Small reproducible probes for attention behavior, token attributions, and model failure modes.
Technical communication piece explaining sequence alignment, scoring, and search intuition.
Notebook bench for baselines, feature engineering, cross-validation, and metric-driven iteration.
A cleaner pass through education, internships, and applied ML practice.
EduSkills Foundation
Production-minded AI deployment, automation workflows, and MLOps practice across cloud-backed model pipelines.
EduSkills Foundation
Completed AWS Academy tracks in ML foundations, generative AI, and NLP with hands-on model training.
Cognifyz Technologies
Built TensorFlow deep learning workflows and used transfer learning to improve training efficiency.
EduSkills Foundation
Worked through predictive analytics, regression, classification, clustering, and model evaluation.
EduSkills Foundation
Built and optimized TensorFlow models while strengthening practical deep learning fundamentals.
Awards, hackathon energy, certifications, and recognitions without turning the page into a trophy wall.
Recognition at SRMIST for technical learning and project momentum.
Competitive programming and Python-focused technical participation.
Deep learning, ML foundations, generative AI, and NLP coursework.
Data engineering, analytics, and applied ML practice across guided programs.
Compact focus areas I keep returning to while building.
Attention behavior, token dynamics, and practical model debugging.
Tool-using systems that can plan, remember, act, and explain decisions.
Small experiments that make neural models easier to inspect and compare.
Perception, feedback loops, and edge intelligence for messy environments.
A terminal-style interface inspired by how I think about systems, AI, and building things.
[01] booting farzan.exe
[02] loading transformer interpretability modules
[03] syncing Kaggle experiment shelf
[04] calibrating autonomous sensing concepts
[05] warning: overengineering tendency detected
[06] gym.exe running in background
[07] ready for recruiter input
[08] listener online_
Open to
AI/ML engineering, data science, applied research, and model deployment.
Transformers, interpretability, autonomous systems, and human-AI workflows.
Student teams, hackathons, prototype builds, and technical presentations.
Projects where clean baselines, useful metrics, and deployable systems matter.
For internships, AI projects, research collaborations, or thoughtful model interpretability work.
Based in Chennai, Tamil Nadu, India. Available for sharp ML problems, thoughtful AI work, and technical experiments with real constraints.