Bridging the gap between bleeding-edge research and production-grade AI. I architect scalable systems that see, understand, and solve the unsolvable.
Choose your experience:
A narrative journey through challenges, solutions, and the “Why" behind the code.
A concise, data-driven summary of skills, achievements, and metrics.
Hello, I'm
Let me tell you a story...
I build production-ready ML systems that solve "unsolvable" computer vision challenges and deliver massive business impact. My focus is on bridging the gap between bleeding-edge research and scalable, user-loved products.
Case in point: I single-handedly re-architected an ML pipeline to be 6x faster, slashing infrastructure costs by over $1 Million (~95%). I also pioneered an industry-first computer vision system to deliver a company's #1 most-requested feature, developing novel techniques to solve a challenge no competitor could.
From designing novel evaluation metrics for autonomous vehicles [IROS '25] to leading teams deploying analytics across 20,000+ cameras, I specialize in turning complex technical problems into robust, high-performance solutions.
Years Exp.
Citations
Boston University
I've always been fascinated by the “black box" of Artificial Intelligence. But for me, the magic isn't just in the mathematics—it's in the application. It's about taking a theoretical paper that says “this is possible" and engineering a system that says “this is profitable, scalable, and real."
My journey has been defined by a refusal to accept “that's impossible." Whether it was slashing infrastructure costs by 95% when everyone said we needed more servers, or building a computer vision system for wildlife that can see in the dark, I thrive where research meets the rigorous demands of the real world.
I am a Machine Learning Engineer, a Researcher, and a Builder. I don't just train models; I architect the future of how machines perceive our world.
My Professional Journey
Fresh out of university, I didn't just join a team; I was tasked with building one. As a Computer Vision Lead at Wobot.ai, I found myself orchestrating a symphony of 20,000+ cameras. The challenge? Real-time analytics at a scale that breaks standard pipelines.
I led a team of 14, and together we built “WoUtils"—a core library that became the backbone of our engineering. We reduced false positives, optimized inference, and saved the company 50% in development time. It was my boot camp in high-stakes, production-grade AI.
Hungry to push the boundaries of what I knew, I moved to Boston to pursue my Master's. Here, at the H2X and BIT Labs, I dove into the esoteric world of Autonomous Driving and Generative AI.
I tackled the “Sim2Real" gap—the notorious difficulty of training robots in simulation and having them work in the real world. My research, which developed novel metrics for evaluating autonomous safety, was accepted into IROS 2025. It was a validation that my work could stand on the global stage of robotics research.
Today, at Moultrie, I face my biggest challenges yet. When I arrived, they had a massive infrastructure bill and a “wish list" of features deemed too difficult to build.
I took a Databricks system that was bleeding money and re-architected it using NVIDIA Triton. The result? A $1 Million annual saving. But I didn't stop there. I built an industry-first “Animal Re-identification" system and a “Night Image Enhance" feature that competitors are still trying to figure out. I turned the “unsolvable" into the “deployed."
Innovation & Engineering
Generate High Quality 4K Wallpapers from Simple Prompts using Stable Diffusion and Image Enhancement techniques.
End-to-end Conditional Imitation Learning framework in a Real-World model city. Focused on safety-critical scenarios.
Generate 3D renderings of an appearance edited object through text prompts using Neural Radiance Fields (NeRF).
FCN based Background Subtractor to extract unseen foreground objects using deep autoencoders.
Privacy preservation algorithm using Intel OpenVINO Face Detection, optimized for real-time CPU performance.
YOLOv3 based object detection to capture Helmetless Riders and their License Plates. Used synthetic data generation.
Every project is a question I needed to answer. Can we generate 4K wallpapers from thin air? Can we teach a car to drive safely in a chaotic city? Here are the answers I found.
Wallpaper AI was born out of curiosity. With the explosion of Stable Diffusion, I wanted to see if I could create a tool that didn't just generate images, but created art suitable for high-resolution displays.
I engineered a pipeline that takes simple prompts and upscales them into breathtaking 4K wallpapers. It's a testament to the power of modern Generative AI when tamed by careful engineering.
Try It Out
Autonomous Driving is the ultimate computer vision challenge. In this project, I moved away from simple lane following to “Conditional Imitation Learning."
We built a real-world model city and trained a vehicle to navigate complex intersections and safety-critical scenarios. This wasn't just code; it was robotics, hardware, and deep learning working in perfect harmony.
Read the PaperAnd there is so much more...
Publications & Contributions
IROS 2025, arXiv preprint arXiv:2510.08571, 2025
Boston University, 2024
Emerging Research in Electronics, Computer Science and Technology, 2019
International Conference on Information Processing (ICINPRO), 2019
Tools & Technologies
Community Leadership, Awards & Service
Co-hosting the Boston Computer Vision AIR (AI, Autonomy & Robotics) meetup group for the past 2 years. I organize monthly events connecting 50-90+ participants, ranging from graduate students to industry professionals and professors from the greater Boston area. My role involves curating speakers, managing logistics, and fostering a collaborative environment for networking and knowledge exchange in Computer Vision and Robotics.
Endorsements
"He independently comes up with brilliant solutions to challenging research problems, while always keeping up with recent advancements in AI and computer vision."
"He exceeded our expectations by far in what he was able to accomplish. We were able to accomplish multiple ML/AI goals this summer thanks to Animikh's efforts."
"'Genuine expert' is the phrase that pops into my mind when I think about Animikh. I've never seen anyone handling multiple projects like him."
"Animikh defined his own research project within end-to-end autonomous driving systems and holistically tackled it... The defense was outstanding."
"He demonstrated exceptional research, analysis, and development skills that greatly contributed to our projects."
"Animikh is one of the hardest-working people I have ever met... Whenever something is assigned to him, rest assured it will be completed with utmost dedication."
"Animikh is one of the most focused and driven people I've worked with... He's a tech whiz, a great leader, and a fantastic coworker."
"Animikh doesn't only know how to deliver, he knows how to deliver well. He has contributed greatly to Wobot's goal of automating and scaling AI processing."
"He has carried out several projects and has published papers under my guidance... he has shown exceptional growth, dedication, and interest in Machine Learning."
"He has got exceptional skills when it comes to coding and research. Animikh has been a great mentor for the Computer Vision Engineers and Interns."
Let's Build The Future
Boston, MA 02134
