I build machine learning systems, AI applications, and developer tools across the full stack. My interests include large language models, agentic AI systems, retrieval-augmented generation (RAG), reinforcement learning, and scalable backend infrastructure.
I enjoy taking projects from research and experimentation to deployable applications, with a focus on performance, reliability, and user experience.
I am an AI/ML Engineer and full-stack developer with hands-on experience building production AI systems from scratch. I turn research into working products that real people use. My work spans RAG assistants, real-time voice agents, GPU-optimized world models, and LLM alignment pipelines.
I thrive in customer-facing, ambiguous environments and take ownership of outcomes from discovery to deployment. I am currently pursuing my Master's in Computer Science at Felician University.
Deconstructed Transformer and Hierarchical Risk Model (HRM) architectures using PyTorch to analyze layer-wise attention weights, improving model interpretability for high-dimensional financial and scientific datasets.
Engineered automated visualization pipelines within a Graph-based IDE using Deep Graph Library (DGL) to track complex data lineage and multi-dimensional system dependencies.
Manage campus IT helpdesk operations, supervising a team of 4 student technicians.
Optimized internal knowledge base and incident workflows, cutting average ticket resolution time by 15% across the university.
Maintain 95%+ user satisfaction rate through consistent service quality and rapid issue resolution.
Audit and maintain campus computer labs, deploy software updates, and manage enterprise licensing.
Led a cross-functional team of 4 developers to build, test, and deploy a secure mobile application using React Native, cutting the launch timeline by 30% using Agile sprints.
Developed a Bone Cancer Detection model using Convolutional Neural Networks (CNNs), achieving a 15% increase in diagnostic accuracy over legacy baseline models.
Engineered a high-throughput data preprocessing pipeline, reducing medical image analysis time by 30%.
Built the community, started podcast, collaborated with companies, managed events.
Collaborated with communities such as Kubernetes, organized events like GDSC WOW and Cloud Native Days, participated in Google Cloud Jam, delivered a guest lecture on blockchain in Andhra Pradesh, and took part in hackathons.
Led a team of three to develop user and admin apps for managing the Texephyr event.
SFT + DPO on Apple Silicon
Developed an end-to-end alignment pipeline for fine-tuning LLaMA models using Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO), optimized for Apple Silicon hardware.
Dreamer-Style Reinforcement Learning
Implemented a Dreamer-inspired Recurrent State Space Model (RSSM) trained on pixel observations. 5M-parameter model with 27.4x training acceleration using Apple Silicon MPS.
Research Assistant
Built a retrieval-augmented research assistant capable of answering questions from document collections while generating interactive knowledge graphs from extracted entities.
Conversational AI
Developed a real-time conversational voice agent supporting bidirectional streaming, tool invocation, interruption handling, and telephony integration with Twilio.
Educational & Research
Implemented GPT-style transformer architectures and sequence-to-sequence models from first principles to deepen understanding of attention mechanisms and autoregressive modeling.
A collection of advanced deep learning architectures and algorithms built from scratch.
Implementation of Research Paper
Implementations of various Transformer architectures including Multi-headed attention, Transformer XL, GPT, MLP-Mixer, ViT, and Switch Transformer.
Implementation of Research Paper
Implementation of Recurrent Highway Networks with enhanced depth and sequential processing capabilities.
Implementation of Research Paper
Deep learning models utilizing Long Short-Term Memory networks for processing sequential data.
Implementation of Research Paper
Implementation of HyperLSTM - utilizing a smaller network to generate weights for a larger LSTM network.
Implementation of Research Paper
Implementation of Residual Networks to train extremely deep neural networks via shortcut connections.
Implementation of Research Paper
Implementation of ConvMixer, substituting convolutions for self-attention and MLP operations in vision tasks.
Implementation of Research Paper
Implementation of Capsule Networks to better model hierarchical relationships in image classification.
Implementation of Research Paper
Implementations of GAN architectures including Original GAN, Deep Convolutional GAN, Cycle GAN, Wasserstein GAN, and StyleGAN 2.
Implementation of Research Paper
Implementations of Generative Diffusion models including Denoising Diffusion Probabilistic Models (DDPM).
Implementation of Research Paper
Implementation of Sketch RNN for generating vector-based drawings using seq2seq VAEs.
Implementation of Research Paper
Implementations of Graph Attention Networks (GAT) and Graph Attention Networks v2 (GATv2).
Implementation of Research Paper
Solving games with incomplete information, such as Kuhn Poker, using Counterfactual Regret Minimization (CFR).
Implementation of Research Paper
Implementations of RL algorithms like PPO, Deep Q Networks, Prioritized Replay, and Dueling Networks.
Implementation of Research Paper
Implementation of deep learning optimizers including Adam, AMSGrad, Adam with warmup, Noam, Rectified Adam, and AdaBelief.
Implementation of Research Paper
Implementations of Batch, Layer, Instance, Group, Batch-Channel Normalizations, and Weight Standardization.
Implementation of Research Paper
Implementation of Knowledge Distillation techniques to transfer knowledge to efficient models.
Implementation of Research Paper
Implementation of Adaptive Computation models like PonderNet to dynamically adjust computation steps.
Implementation of Research Paper
Utilizing Evidential Deep Learning to thoroughly quantify classification uncertainty in neural networks.
Relevant Coursework: Data Science, Data Mining, Artificial Intelligence
Relevant Coursework: Deep Learning, Linear Algebra, Calculus
My latest technical writings and thoughts published on Substack.
Fetching latest articles...
Exploring the intersection of AI, Quantum Computing, and Philosophy on Cyber Socratic.
The Socratic Embers
A deep dive with Kristofer A exploring the evolution from low-level systems programming to modern artificial intelligence paradigms.
The Socratic Embers
Discussion with Jeff Smith, founding leader at PyTorch, covering Open Source and Startup Innovation in the AI landscape.
The Socratic Embers
Kevin Talcott shares insights integrating economic strategy with tech philosophy for scaling sustainable success.
The Socratic Embers
Discussing the mental frameworks needed to build clarity and focus amid rapid technological distractions.
DeepLearning.AI
Issued Feb 2025
I am interested in opportunities related to machine learning engineering, AI infrastructure, agentic systems, full-stack AI development, and forward-deployed engineering.