FIG 1.0: PROFILE_VIEW
Suraj Mittal
Engineering Manager
"If I try my best and fail, well I've tried my best" - Steve Jobs. I am passionate about new technologies and always think about the changes these technologies bring in this ever changing world for the better of it. Currently leading the platform backend at Setu.
SPECIFICATIONS
SYS_HISTORY_LOG
Setu
Engineering Manager
April 2025 - PresentLeading engineering teams at Setu. Building Finstack to democratise finance.
Staff Software Engineer
April 2024 - March 2025Leading technical initiatives and architectural decisions.
Senior Software Engineer
March 2022 - March 2024Core contributor to the payments and infrastructure stack.
Artefact (Client: Morpheus LLC)
Solutions Architect
Owned discovery of client requirements across big data and web stack initiatives, proposing and delivering end-to-end solutions.
Rivi
Technical Advisor
Feb 2021 - March 2022Guided the tech team and built hiring procedures for new technical hires.
Technical Architect / Senior Data Scientist
July 2019 - Feb 2021Built recommendation and optimization systems, a booking chatbot, and ML service backends; automated deployments with Kubernetes, monitoring, and a GitLab CI/CD migration.
EdGE Networks Pvt. Ltd.
Senior Data Scientist
Built knowledge graphs and language models, improving skill coverage to 90%, boosting demand analysis by 10% with KG embeddings, and increasing parsing performance by 20% via entity normalization.
Hashworks IT Services Pvt Ltd
Senior Machine Learning Engineer / Data Scientist
Built the AI core for a recruitment platform, including ranking and recommendation, skill suggestions, a job crawler on serverless infrastructure, and a resume parser.
Freecharge
Software Engineer
Designed and developed the Windows 10 UWP application.
Kirusa
Associate Software Engineer
Built the Instavoice Windows 10 UWP application and an Amazon Alexa skill.
ACADEMIC_RECORDS
Visvesvaraya Technological University
Bachelor's Degree, Computer Science
PROJECT_SCHEMATICS
Particle Swarm Optimization as Meta-Heuristics for LSTMs
Showed that combining PSO with constrained-boundary LSTMs improves performance and reduces training time versus vanilla LSTMs.
Sparse Document Composition Vectors
Built an unsupervised document representation method using soft clustering, word2vec, and threshold-based cutoffs for downstream tasks.