Hi, I'm Sk Jishan, a Computer Science undergraduate at IIIT Kalyani passionate about Software Engineering, AI/ML, and Generative AI.
I have hands-on experience building full-stack AI applications using Python, React, Django, LLMs, RAG, and LangChain. I enjoy solving complex problems and have solved 3000+ coding challenges across platforms, achieving a LeetCode rating of 2163 (Top 1.11%) and Codeforces Expert status.
I'm always eager to learn, build impactful products, and explore cutting-edge technologies in software engineering and AI systems.
Pinned Loading
-
AI-PM-Assistant-using-RAG
AI-PM-Assistant-using-RAG PublicAI-powered RAG assistant for Product Managers β ask natural-language questions across PRDs, meeting notes, and Jira exports, with grounded, source-attributed answers.
Python 1
-
hybrid-cnn-vit-medical
hybrid-cnn-vit-medical PublicHybrid CNN + Vision Transformer for skin lesion classification on HAM10000. Cross-attention fusion combining ResNet's local feature extraction with ViT's global context. Includes ResNet50 and ViT-Sβ¦
Python 1
-
pinn-climate
pinn-climate PublicPhysics-informed ConvLSTM for satellite climate prediction. Constrains deep learning with advection-diffusion PDE loss + adaptive Ξ» weighting β validated with an honest ablation study on real NOAA β¦
Python 1
-
dpo-tool-calling
dpo-tool-calling PublicDPO for Tool/Function Calling - Fine-tune LLMs to intelligently decide when to call external APIs vs generate text using Direct Preference Optimization. Achieve 94.1% schema accuracy on Llama-2-7B β¦
Python 2
-
dynamic-depth-transformer
dynamic-depth-transformer PublicToken-level early-exit Transformer with Gumbel-Softmax routing and a latency-aware joint loss β skips redundant layers for "easy" tokens while going deep on hard ones.
Python 1
-
domain-adaptive-llm-finetuning
domain-adaptive-llm-finetuning PublicTwo-stage CPT + SFT domain adaptation pipeline (QLoRA/Unsloth) with an evaluation harness measuring catastrophic forgetting via MMLU/GSM8K retention vs. domain-accuracy gains. FastAPI serving + CI/CD.
Python 1
If the problem persists, check the GitHub status page or contact support.

