SFL Scientific: Amalgamating People and Technology

Follow SFL Scientific on :

Michael Segala, Co-Founder & CEO, SFL ScientificMichael Segala, Co-Founder & CEO Since rising to prominence in late 1990s to hitting the headlines in 1997 for winning against a chess expert to evolving into Siri, Google Now, Amazon Echo, and the like, virtual assistants have emerged as force to reckon with. The rise of natural language processing (NLP) in the backdrop has planted these assistants as an integral part of workplace communication. While interfaces have long been imagined strictly in a visual sense, the advances in NLP are changing the scenario by powering everything from chatbots to language translation and voice-activated platforms.

Virtual assistants are a hallmark of NLP technologies, but unfortunately, only a third of smartphone owners today use them on a regular basis. The users have the luxury of vocally asking their devices through this pragmatic technology to either search for an item or perform an action. You could request a virtual assistant to “Find shopping malls in Boston” and then ask a follow-up question “How about near Cambridge?” without mentioning shopping malls for a second time. With NLP being applied to support conversations between smartphones and their owners, the concept of human-machine communication, also known as computational linguistics, has reached a whole new level. Seizing this advancement as an opportunity to drive innovation, SFL Scientific, a boutique data sciences consulting and professional services firm, marks its niche in uncovering business value for its clients through cutting-edge NLP solutions.

The Journey so Far

SFL Scientific has paved its own path through the NLP arena by consolidating a pool of best data science and analytics talents to deliver an unparalleled level of service with added value and ROI. “We believe strongly in fielding teams with complementary skills and deep expertise that use a holistic approach to data science,” states Michael Segala, the co-founder and CEO of SFL Scientific. “And we always seek to learn from the AI community, our clients, and our employees to deliver a competitive advantage with custom solutions, and magnify our efforts and continue to transform organizations.” From hardware to algorithm implementations to securing higher business values and procuring quicker results, SFL Scientific has established a unique culture to bring the most relevant and complete answers to a business situation. With rich expertise in data engineering, artificial intelligence, big data, and machine learning, the firm embarks on a mission to develop and implement comprehensive data strategies and mining massive datasets. Having identified the client NLP capability gaps to uncover business value through predictive analytics, SFL Scientific combines its deep sector expertise with agile and pragmatic approaches to change that narrative.

"We believe strongly in fielding teams with complementary skills and deep expertise that use a holistic approach to data science"

The Helping Hand

SFL Scientific draws on its specific domain knowledge and fundamental understanding of core business requirements to offer reliable data science and custom developed NLP solutions for sophisticated R&D type problems. The industry agnostic firm employs a holistic approach by combining the latest technological advances with real-world expertise and adds a dash of innovation to solve a wide range of enterprise-scale data-intensive challenges.

Achieving AWS ML Competency status recognizes our proven track record of delivering advanced solutions in data science, automation, and machine learning


With its arsenal boasting of cutting-edge NLP techniques to weed out the noise, SFL Scientific helps clients thrive in today’s rapidly changing information landscape, and generate tangible and operational value. The team at SFL Scientific leverages its competency in deep learning, NLP, machine learning, and predictive analytics to craft robust, automated tools and end-to-end pipelines that deliver actionable insights, and drive growth pertinent to machine intelligence. In addition to that, the firm imparts profound technical knowledge of NLP, machine vision, big data analytics, time series, and data warehousing along with lean project management to ensure clients achieve successful last-mile development. That’s not all, SFL Scientific goes beyond delivering customized NLP solutions by capitalizing on its proficiency in machine learning algorithms, data mining techniques, and business cases for assisting clients to develop new technology and drive improved revenue. When a client’s data-driven business is faced with a predicament to reap the benefits of user-generated data and maximize ROI, SFL Scientific as a torch bearer leads the way, from project inception to NLP integration and deployment.

Data Consulting: What and How?

The data consultants, scientists and engineers at SFL Scientific combine efforts to analyze business processes and pinpoint potential areas for restructuring or improvements. With the amount of user-generated data growing at unprecedented rates, the firm assists enterprises in dire need of predictive analytics and automation to drive business value from the information accumulated. Sure, AI immensely simplifies the predictive analytics and automation aspects, but the key to reshaping businesses lies in freeing up time for innovative and strategic initiatives. Starting from understanding business use cases and implementing comprehensive data strategies to leading projects and performing extensive R&D, SFL Scientific develops and provides customized NLP solutions best-suited for its clients. Rather than pushing a one-size-fits-all approach, the firm tailors its solution to address specific problems, empowering clients with the flexibility to transform current workflows and automate business processes. As a data scientist’s effort is not deemed complete without any actionable business feedback, SFL Scientific enables organizations to consume user-generated data in a predictive and prescriptive manner, and accomplish their project goals.

While every facet of a business is virtually open to data collection, the team at SFL Scientific believes that data revolution spans four main types of information: text, image, time-series, and consumer and company data. The firm then utilizes relevant data to converge on a multitude of business outcomes, thus increasing the speed to market, sales, reduce costs with improved quality of service. SFL Scientific takes it upon itself to employ a fine-tuned data strategy approach for building appropriate data science and NLP solution while efficiently meets the clients’ unique business goals. As a full service and data consulting firm, SFL Scientific applies a research and development method to investigate the best possible outcome for every data problem and execute the vision at the highest level. From examining data sources and shaping information to evaluating and creating an effective NLP model for delivering optimal results and insights, SLF Scientific seamlessly performs it all.
That said, SFL Scientific designs NLP systems that learn, augment human decision making, and generate reliable insights, which allows clients to utilize their data, operate, and innovate and develop new products efficiently.

Reaping the Benefits

The prowess of SFL Scientific’s customized NLP solution was put to test when LinkSquares—a contract analysis and reporting service provider—approached the firm to algorithmically extract key terms from legal contracts and documents into their Smart Values contract analysis cloud. As time is of utmost importance for legal and finance departments, SFL Scientific automated the entire procedure of reviewing contracts with NLP, which immensely simplified the information extraction process. The extraction process via NLP comprised of three main steps: feature engineering, model stacking ensemble, and post-processing. SFL Scientific deployed a customized NLP algorithm through AWS that ran the code to extract key terms on demand whenever a legal document was uploaded. “Achieving AWS ML Competency status recognizes our proven track record of delivering advanced solutions in data science, automation, and machine learning,” adds Segala. “Our team is dedicated to helping organizations, many of whom utilize AWS to achieve their business goals through data-driven initiatives.”

Once the key terms were extracted from a legal document, the next step was to classify raw tokenized text into pre-defined categories. Each word in the document was tokenized using a regular expression tokenizer, parsed, and stored as an independent observation. SFL Scientific implemented a model stacking ensemble technique to better predict the class of individual tokens. Following which, the firm deployed XGBoost, a gradient boosted decision tree-based model as a meta-classifier to finalize and clean the class predictions. These classified token were continuously concatenated with each other to produce a more homogenous output.

The NLP algorithms generated significantly superior classification results than the models using F-measure as the scoring metric. Backed by AWS’ powerful computing capabilities, SFL Scientific’s NLP algorithm streamlined the token classification operation and made this asset easily accessible to all LinkSquare employees. Impressed by its efficiency of, LinkSquares deployed SFL Scientific’s NLP algorithm to fruitfully analyze over 100,000 legal contracts and documents in a streamlined, transparent, and quantitative manner. “For over a year, we’ve developed a strong relationship with SFL Scientific and leveraged their skills to develop and deploy machine learning in our systems,” says Eric Alexander, CTO of LinkSquares CTO.

The Road Ahead

As SFL Scientific looks to add more such success stories to its triumphant track record, the firm’s management team capitalizes on its experience in high-tech consulting and analytic modeling to create actionable and lasting impact for its clients. Under their invaluable guidance and leadership, SFL Scientific aims to harness its theoretical, computational, and analytical expertise to deliver robust data analysis and support for clients’ R&D and production efforts. With curiosity, efficiency, and flexibility at its core, the firm intends to provide tailored and flexible NLP solutions custom to the clients’ business goals along with an excellent standard of service. Moving forward, SFL Scientific aims to maintain unparalleled professional standards, and a central focus on guiding clients through project roadmaps, milestones, and data strategy sessions for uncovering dependable, long term business value. “We’re excited to further our ability to educate customers, provide return on hardware, and leverage their data as we drive value through new systems,” concludes Segala.
- Kenneth Thomas
    July 17, 2019