Sanjana Sahayaraj

I'm

About

Experienced working across research, solutioning, engineering and operations lifecycles of industry grade AI / ML products that serve business clients (B2B) and end users (B2C).

Data Scientist, Analytics Engineer and ML Engineer

Worked across:

  • Capacity:  Individual contributor and Tech Lead
  • Organization Types:  Multinational company and startups
  • Contract Types:  Full Time, Contract and Freelance

  • Education: B.E. in Computer Science (University First Rank - Gold Medal), M.Sc. in Computer Science & MBA in Business Analytics
  • Roles held so far: NLP Research Engineer, Senior and Lead Data Scientist, Staff and Principal Data Scientist
  • Institutions: B.E. from SSN - Anna University, M.Sc. from University of California - Santa Barbara, MBA from BITS Pilani
  • Past and current companies: IBM Almaden Research Center, Kyndryl (Spin-off from IBM), Epifi (Fi Neo Bank), Smarteeva and Endeavor Labs

I finished my undergrad in Science and Engineering in India and Master of Science in the USA, which laid my technical foundations in Computer Science and Machine Learning. I started working full time as an NLP Research Engineer in IBM Almaden Research Center, during which I was involved in the research side of ML and NLP. Followed by this role, I pivoted more into product oriented roles as a Data Scientist and Senior Data Scientist within another business unit in IBM which spun off into Kyndryl. The birth of this new entity, made me realize how I enjoy imparting structure and progress to complex situations and how I derive inspiration from wearing multiple hats within the same job role. Since then, I've specialized in being a founding member of new Data and ML products and I take pride in taking them from scratch to scale.

Recommendations

Here are recommendations from people I've worked closely with, during my career. The same can be found on my LinkedIn profile as well.

Sanjana and I worked together as a two-person team on a research project concerning the merging of data from medical ontologies, and deriving a new kind of word embedding for terms in the ontologies. We worked together on conceptual aspects, on programming, on devising and performing performance tests, and on writing a conference submission describing our work. Throughout I was impressed by Sanjana's initiative, skills, creativity, and teamwork. Her work on combining ontologies using ontology keyword queries as a first step, facility with Python and Github, and writing the majority of the submission manuscript were particularly important. I was also impressed by Sanjana's presentations and contributions to discussions, for a journal club we were both in, studying neural nets for NLP. Sanjana also took initiative in organizing events and activities that helped our larger organization at IBM, before and during the Covid-19 pandemic. This played an important role in keeping our group coherent and sane. It was a pleasure to work with Sanjana, I strongly recommend her.

Sanjana worked on my team as a Software Engineering Researcher at IBM Almaden. I have found Sanjana to have a very strong work ethic which is something I value above everything else. Sanjana is smart, sharp, quick learner and very professional. She works well with people, eager to take on hard problems, follows-through on issues and resolves them. She can multi-task very well and I have seen her juggle many projects and problems while on my team. She is a pleasure to work with. She is moving to another role as a Data Scientist in IBM India. I wish her the very best in all her future endeavors and hope our paths cross again. She will do well in any team she joins and they are lucky to have her. Best Wishes Sanjana!!!

I was lucky to work with Sanjana on a complex NLP prediction problem for several weeks. She impressed me with how quickly she was able to understand the complexity of the issue, her readiness to come up with the right questions, her ability to quickly solve complex tasks while always prioritizing the best solution for the customer. She has a vast knowledge of the NLP domain, ranging from standard algorithms to the latest developments in Transformers and LLMs in general. During the lifetime of the project, she went from developing the initial model to supervise and contribute to all other activities, from data sourcing, to transformation, modeling and productization. She showed the ability to follow the entire lifecycle of a data science project, incorporate the customer needs into it and drive its outcomes towards the best solution for the customer. Working with Sanjana was a great professional experience, she is easy to work with, always happy to help and collaborate on complex tasks and very organized. I learnt a lot from her and anybody would be lucky to have her as a colleague.

Sanjana is our senior data scientist and she has done a great job setting up our environments and data pipelines. We went from having an unreliable and messy data pipeline to a clean, working, and fast data warehouse. On top of that, Sanjana provided nice prediction algorithms and models which made a difference to what Smarteeva was trying to achieve and took the company a long way to our finished product. Sanjana is smart, easy to work with, and very reliable. When something was assigned to her, I knew she would get it done and she would communicate the entire time. I always had confidence in her. Sanjana is a gem and she has a very bright future.

I was Sanjana’s mentor when she worked at IBM Almaden Research Center. We have regular lunch mentoring sessions which we usually discuss career planning, tips to help her grow internal and external eminence, etc. She always prepared for each session with well-thought and articulated questions. I really enjoyed the opportunity to interact with such a young, energetic and dedicated research software engineer. I was always inspired by her capability to connect with members across teams and organizations, her enthusiasm to share her expertise in NLP and help others in many ways. I hope she will enjoy her next step of career in India. I highly recommend her!

Sanjana is amazing at what she does especially in the domain of NLU. She has contributed to the NLU-related problem statement at Fi. Her solutions are directly integrated into the product, creating a better user experience for our users. She's a very quick learner and implements solutions efficiently which was evident by how she learned graph-based modeling and used it in building the solution. From a cultural perspective, she is an amazing fit for our team. She takes a ton of initiatives to improve the existing process, share feedback, and mentor those around her. She's always available for helping and brainstorming problem statements.

Sanjana is a great collaborator, innovator, and technical guru. I worked with her on a Data Science team at IBM creating solutions in cloud computing management. Sanjana not only has a deep technical knowledge but also the unique talent of understanding and being able to push product vision. I was continually impressed by her elegant solutions and knowledge of new technologies and academic research. Sanjana was the biggest driver of innovation in our team. She mentored us in patent writing and encouraged team members to explore many invention ideas. Sanjana is an excellent team member and exceptional Data Scientist, and I would be happy to work with her anytime.

Sanjana is an incredibly bright and versatile data scientist, who brings a powerful combination of machine learning knowledge, engineering skill, and organizational ability. She played a team lead role on one of our client engagements at Endeavor Labs, and did an impressive job coordinating the efforts of a 4-person project team, while also playing the role of senior technical contributor. Any data science or ML team will be lucky to have her on board!

Sanjana and I were part of the same team which was focussed on delivering novel health informatics solutions to the Watson Health Imaging business unit at IBM. We worked very closely to come up with a solution to adapt an existing product to work on EMR in French. Right from the get-go, I realized that Sanjana was a fast learner. She asked the right questions and if something wasn't clear she made sure she had the information that she needed. Her documentation skills, which is something I value very much in a researcher, is second to none. She was on top of her deadlines as well. She is an excellent team player as well and was always there if I needed help or needed to clarify something with her. She will be a valuable asset no matter where she goes. It was real pleasure to work with her.

Sanjana and I worked in the same group and collaborated on a project involving crowdsourcing and Semantic Role Labeling. Her knowledge of NLP and top software engineering skills became apparent during this collaboration. She is eager to share her knowledge. I mentored Sanjana on patenting at IBM and its process. She was a quick learner and went on to officially become an inventor by filing her first patent. She is curious, engaged and creative so I expect her to remain active as an inventor. In addition, Sanjana invests time and effort participating and leading activities that promote community building, team spirit and technical exchanges. I highly recommend her.

I was impressed by Sanjana's ability to handle any situation calmly and patiently, even in the toughest situations. She has proven great experience in the projects that she handles. Her ability to convert the ideas into innovations is truly amazing.She is a great mentor and when it comes to sharing the knowledge, she is a great presenter.

Sanjana is a young, energetic researcher and engineer. I got to know her while we're in (two) same group in IBM Research. She is curious and likes to learn. I enjoyed many technical discussion we had for various problems. She is smart and very motivated. Also, she is a great communicator. I am impressed by her ability to connect people in various departments and make things happen through collaborations. No doubt that she will make positive impacts on her new role in IBM. All the best, Sanjana.

I met Sanjana as part of the Developer JumpStart program at IBM. I was very impressed by her leadership skills, curiosity, and interest to learn new things early in her career at IBM. She is a knowledgeable data scientist and a great speaker. It's a joy to be around Sanjana, and I highly recommend her.

Sanjana is a brilliant data scientist, highly intelligent and thorough. Her ability to break down complex problems helps her to deal with ambiguous situations. She has great communication skills to articulate complex problems and situations. Above all she is a great person who is fun to work with.

Sanjana, you have been truly amazing at all you do, you are excellence beyond par, I like your growth mindset, proactive approach and attitude to go beyond the job and perform in multiple initiatives also build on your technical vitality I was truly amazed at your ability and skill to convert ideas and put them into patents so very quickly That you have been on lecture series on patenting and multiple topics on hyperscalers speaks a lot on your abilities You have added quality to everything we have touched and deserve this award truly and many more to come your way Your MBA degree will further enhance and broaden your perspective and make you a more proficient knowledge specialist. Good Luck !!

Sanjana is a highly dedicated and proficient senior data scientist. She is a very quick learner and well versed in model development for supporting diverse features in multi-cloud management platform. Her application of model building skills is unique and she has researched and built high performing cost optimization models that works well in production. In addition, she is a solid contributor of patents and is a role model for others in the team to pursue. Only with her guidance and support could I submit my first invention work. Being a high end team player she stands out and plays a major role in case of any presentations and infonaut sessions to wider and delivery teams. She always proved to be a super enthusiastic person with ability to pick up any task and deliver with high quality and on time. One such example is her contribution in the MLOPS work including model customization, model serving and monitoring. Data Science being a fast growing and demanding industry sector, I am confident that person like Sanjana will be able to satisfy the needs with her strategic decision-making and technology skills and foster the growth of the project and the organization

I worked with Sanjana developing a real-time incident resolution system for AIOps. Sanjana has great expertise in the subject matter of NLP and was always willing to help other team members beyond her responsibilities, which proved to be a great catalyst in the success of our project. Moreover, she is well organized, diligent, and a fast learner. All these attributes were critical in meeting the deadlines and KPIs.

Sanjana is future vision NLP enthusiastic and innovative senior data scientist. She has great leadership initiative and quick learning abilities to take up any challenge provide agile solutions at scale, she will definitely brings lots of innovative AI/ML driven solutions with her immense knowledge in area of AI/ ML.

Career

I talk about my career progression and milestones under each of these clickable tiles:

Click Here
Click Here
Click Here
Click Here
Click Here

Patents and Publications

Patents

Granted and Published

  • US11200883B2: Implementing a domain adaptive semantic role labeler
  • US20220284996A1: Combining domain-specific ontologies for language processing
  • US20220215325A1: Real Time Identification of Changed-Induced Incidents
  • US20220180068A1: Bias Identification and Correction in Text Documents
  • US20240256916A1: Continuous maintenance of model explainability
  • P202103424US01: Recruitment Augmentation with digital footprint utilization
  • P202103930US01: Automated Customized Machine Learning Validation Flow

Patent(s) Filed

  • P202400066: Prompt Optimizer Aparatus for Large Language Models

Defensive Publication(s)

  • P202100615: Data Management methods for improved model continuity in MLOps
  • P202100886: Literature consolidation through relative text processing on pivot document
  • P202100897: Named Entity Guided Domain Ontology for structured semantic analysis of ITOps text
  • P202102040: Using Model Simulation, Optimization, and Profiling to Generate and Forecast Cloud Infrastructure Cost Savings
  • P202103451: Automated continuous and adaptive knowledge extraction and transfer mechanisms
  • P202103451: Continuous improvement of market success rate through end-to-end intelligent process management on consulting
  • P202101977: Adaptive Intelligent system and methods of creating UX mockups based on existing themes
  • P202400044: A conversation-guided enterprise knowledge base and planner
  • P202400060: Context-Aware Indoor Environmental Management and Navigation Solution for Storage Facilities
  • P202400064: Integrated Data Attribution and Document Cataloguing for Large Language Model Applications

Inventor Awards

  • Invention Plateau
  • Invention Development Team - Silver Level

Open Source Contributions

  • Finance Proposition Bank (FinProp)
  • Contracts Proposition Bank (ConProp)

Publications

  • Order Embeddings from Merged Ontologies using Sketching - Kenneth L. Clarkson, Sanjana Sahayaraj
  • Knowledge Discovery through Computational Methods on EEG and fMRI Data - Sanjana Sahayaraj, Shomona Gracia Jacob
  • Formal Methods for Business Processes : A Survey - Sanjana Sahayaraj, S Sheerazuddin

Talks

  • SystemT: Declarative Text Understanding for Enterprise - KDD 2019
  • Introduction to and overview of Data Science - IBM CSR at VIIT
  • Exploring NLP Techniques to Help Build Medical Decision Support Systems - Computing Research Association

Blogs

  • Getting Data Science / ML Certified across Azure, AWS and GCP - LinkedIn
  • How to local LLMs - SubStack

Skills

While I'm learning and developing myself as a person and a professional everyday, these are some of the skills I have honed and can use to hit the ground running.

Uncovering hidden potentials in data

Whether it's a useful feature or a hidden risk pattern, I can uncover new paths, both at high level and deep in the weeds through EDA, hypothesis testing and quick modeling POCs.

Bringing structure to abstract problem statements

Every great project starts with a vision and not necessarily an entire plan or path laid out. This is where I can help with my organisational and planning skills to define the path, modules and implementation & validation plan.

End to end model lifecycle ownership

I have established track record of helping business metric growth through continued improvement of ML systems (MLOps) in accordance with evolving business or market needs.

Researching, publishing and patenting

Through my academic background and starting my corporate career in a research lab, I carry the innovative spirit everyday. Additionally, I am also well versed in publishing and/or patenting the same to establish reputation and claim IP.

Bringing the ML product to life and out into the world

I have gone beyond POCs and research version of models, to ship products consumed by both corportate and retail users. And from there, integrated user feedback to ship improved versions.

Mentoring on technical and soft skills

I have been in tech lead and dedicated formal mentor roles coaching across data science, invention drafting and presentation skills.

Outreach




I am a guest lecturer at Birla Institute of Technology's Work Integrated Learning Programme, outside work hours on weekends relating to

  • Data Science
  • Data Engineering
  • MLOps
  • Cloud



I am working on a series of courses on Applied ML on Udemy and Tutorials Point, which are designed using the principle of always tying back theoretical concepts to real world applicative scenarios, it's benefits and limitations to truly get the learners ready to apply their knowledge in the real world, be it in their regular job, venture or a new project




I run a podcast on youtube, spotify and apple podcast for the data and ML community called Your Data HQ (YDHQ) to come together and learn from each other on topics of Data, Analytics, ML research, real projects and life outside work




I am a mentor for Apziva's AI Residency Program where I mentor and coach the professional residents on gaining and applying industry experience and skills, as they work on projects from different industries. Additionally I have been part of SWE (Society of Women Engineers) programs and mentor network since 2019

Contact

My preferred contact methods are through LinkedIn and Email. Do connect if you'd like to connect on something or get a copy of my updated resume.