Perfios - Data Scientist - Deep Learning (2-5 yrs)

About Perfios :

Founded in 2008, Perfios is a global B2B SaaS fintech company serving the Banking, Financial Services and Insurance industry in 18 countries, empowering 1000+ Financial Institutions. Through its pioneering software platforms and products, Perfios helps Financial Institutions take big leaps by shaping their origination, onboarding, decisioning, underwriting and monitoring processes at scale and speed. Perfios delivers 8.2 Billion Data Insights to Financial Institutions every year to facilitate faster decisioning, and processes 1.7 Billion Transactions a year with an AUM of 3 Trillion, significantly accelerating access to credit and financial services for their clients' customers.

Headquartered in Bangalore, Perfios has offices worldwide and with 75+ products and platforms, and over 500+ APIs. In Perfios, its clients have a confidant and a robust start-to-end tech platform.

Funding :

Perfios secured $80 Million from Teachers Venture Growth, the late-stage venture and growth investment arm of the Ontario Teachers Pension Plan in March 2024. Earlier in FY 23-24, $229 Million was raised by Perfios in their Series D round from Kedaara Capital, marking the largest fund raise among Indian B2B SaaS companies for this financial year. Prior to this, Perfios raised $70 Million in Series C round from Warburg Pincus and Bessemer Venture Partners in February 2022. The latest round of funding has taken the total funding raised by Perfios to $435.1 Million.

Key Metrics :

- 75 Million bank statements processed annually

- 3 Trillion in loans processed for Banks & Financial Institutions per year

- 1.5 Trillion API requests processed annually

- 200+ Infosec audits per year

- 0 Infosec violations since inception

- 99.9% Uptime with 0 SLA violations

Current Employees : 1400+

What your average day would look like :

- Collaborate with product and engineering teams to understand requirements and devise possible solutions.

- Explore existing research papers, ideas and codebases that can be leveraged in current tasks.

- Search for open source datasets and/or design synthetic data pipelines (including data augmentation).

- Devise and implement experiments using DL/ML models.

- Evaluate the experiments to find failure patterns and come up with improvements in data/model architecture/loss function etc.

- Communicate results and ideas to key stakeholders.

- Optimize the models for production and collaborate with software engineers for deployment.

Must have skills :

- Hands-on experience in dealing with image data and CNN based architectures

- Should have worked on deep learning frameworks (like pytorch, tensorflow, keras etc.)

- Proficient in Python and packages like Numpy, Pandas, OpenCV

- Good understanding of data structures and algorithms along with OOPS, Git, SDLC

- Mathematical intuition of ML and DL algorithms

- Good understanding of Statistics, Linear Algebra and Calculus

- Should be able to perform thorough model evaluation by creating hypotheses on the basis of statistical analyses

Highly desired :

- Hands on experience with latest computer vision model architectures and concepts like ViTs, GANs, Diffusion, Vision Language Models

- Knowledge of training and inference optimizations using CUDA, C++, ONNX, TensorRT, OpenVino etc. and profiling of ML pipelines

- Worked on building production level APIs for serving models (Flask, Django, TF Serving)

- Hands on experience of using MLOps tools