I am a Research Assistant working at the Institute for Analytics and Data Science at the University of Essex.
Even our beloved ChatGPT knows about me.
I am fascinated by anything related to Mathematics, AI, Quantum Physics, Space and technology in general.My favorite genres for reading and binge-watching are Sci-fi, Geopolitics and Sitcom.
I am a huge Star Trek (TNG) fan and I wish I could comprehend complex subjects and play music like Lieutenant Commander Data did, but my curiosity level certainly can match his!
But Why Lieutenant Commander Data?
Here is a conversation of Data(An android- Artificial Life form) and Timothy(a human who wants to be like commander Data) from
Start Trek The Next Generation, episode Hero Worship [5.11], that accurately describes
how a sentient AI being would feel about humanity.
Data: I have often wished to be human. I study people carefully, in order to more closely approximate human behaviour.
Timothy:Why? We're smarter and stronger than humans, we can do more than they can.
Data:But I cannot take pride in my abilities. I cannot take pleasure in my accomplishments.
Timothy:We never have to feel bad, either.
Data:I would gladly risk feeling bad at times, if it also meant that I could taste my dessert.
© Image Credit: Data and Spot - Star Trek Timelines DataCore
I may not be an expert in AI research yet, but my impactful career as Research Assistant and previously in IT consulting has taught me enough to consider myself a problem solver.
In my experience, a person's aspiration and aptitude play very crucial role for solving any problem and often outweighs the skill-set.
These are list of my skills that may be useful to understand what I am familiar with when it comes to tools and technology!
->CV: CNN,ResNET,GANs,Fast-RCNN, Llama3.2 Vision
->Open Models:Chat, NLI and Embedding models hosted on Huggingface and Ollama
->Closed Models: GPT3.5, GPT4, GPT4o-mini, Claud, Gemini1.5-flash, Gemeni1.5-pro
->Programming: Python
->Data: Pandas, MongoDB, Essbase, Excel
->Scikit-learn, Numpy, Matplotlib
->Deep Learning- PyTorch
->Web: FastAPI, Flask
->Deployments: Docker, Chainlit
I have extensive experience in OLAP data modelling, data integrations, workflow automation and financial data analysis. I consulted several customers from various industries for designing and implementing their planning, budgeting and forecasting business processes using Oracle Hyperion, Essbase and Oracle EPM cloud platforms.
Not an expert but these languages and software that I can understand
and can easily debug codes, do code-reviews and develop applications:
C, Java, Groovy, Jython, MATLAB, SQL, VB, Bash and Powershell
I am working on a Innovate UK sponsored project with Algomo.
This is currently work in progress. More details will be updated soon.
Meanwhile:
This is a computer vision problem from the pharma industry.
When the input is significantly larger and is at microscopic level, the labelling pose a unique set of challenges.
I explored these challenges and a framework developed by Google research, called
The main objective was to find a solution to the ever challenging problem of accurate labelling and the associated cost without losing the effectiveness.
My solution resulted in
This review thoroughly examines the present landscape of explainable artificial intelligence (XAI) tools, placing particular emphasis on recent progress in both model-specific and model-agnostic methodologies.
This presentation was part of my coursework in Digital Signal Processing, where I provided an introductory overview of Speech, discussed components and solutions for Speech enhancements, and a couple of explored existing solutions that address the issue through Deep Learning.
