Mar 2022 – Apr 2024 · 2 yrs 1 mo · Gurugram, Haryana, India · On-site

• ● Help develop a creative atmosphere to innovate, while holding the team accountable for making smart decisions and delivering results.
• ● Requirement Elicitation, Documentation & Testing
• ● Set the technical direction for the Deep Learning and Backend Engineering teams. Lead, motivate, and mentor a team of cross-domain rockstar developers.
• ● Set KPIs and KRAs for the Agile Team and acting Product Owner for GrainScanner and FruitSort
• ● Supported and managed talent acquisition and improved hiring timelines.
• ● Managed multiple stakeholders and coordinated with market heads for pre-sales support.
• ● Designed and developed high speed large scale object detection and instance segmentation on high-resolution images 8000 x 10000, used for quality inspection of spices.
• ● Pioneered innovative high speed 360 degree scanning using CMOS scanners for spices and grains for quality inspection and grading using Instance Segmentation.
• ● Improvised low latency computer vision-based fruit defects segmentation method using differential color spaces and CNNs to run on edge with an accuracy of 97%.

Jul 2021 – Feb 2022 · 7 mos · New Delhi, Delhi, India

• ● Developed the logical architecture in support of the overall vision of the platform, identifying all of
• the required components and subsystems, including their individual responsibilities.
• ● Exploring Triton Inference Pipelines
• ● Requirement Elicitation, Documentation & Testing
• ● Cultivating and growing talent to be capable of achieving outstanding results.

Feb 2020 – Jul 2021 · 1 yr 5 mos · New Delhi, Delhi, India

• ● Designed & Orchestrated Kafka Backlog Based Kubernetes custom Horizontal Pod Autoscaler for
• AI model inference on a GPU resource-limited node.
• ● Develop 2D/3D Keypoint localization using CNN for automatic pose estimation of cars.
• ● Innovated Automatic Driver Assessment System for Smart Yard using multi-camera system to detect
• and track car and assess driving behavior applying visual odometry. Patent Pending.
• ● Deep learning-based Lane Occlusion classification with 91% accuracy validated on real feeds.
• ● Work with cross-domain teams like DevOps/QA/Data Analytics/Annotation Specialist teams on
• technology decisions and technical designs
• ● To help develop backend micro-services for C3 crowdsourced data annotation engine with automatic
• data train-validate-test and model retraining capability with KPI dashboards

Jan 2019 – Feb 2020 · 1 yr 1 mo · Gurgaon, Haryana, India

• Delivered improvements for Video Management System and VLC Player C-API backend and
• reduced overall latency of video trans-coding by 12%.
• Spearheaded solutions engineering and system redesign using Strategy Design Pattern that
• eliminated dependency of rebuilding and integrating VLC backend.
• Implemented & improved existing Tiny Yolo for Real Time Edge Video Analytics on NVIDIA
• Jetson & Intel Myriad, 1.6x Speed-up using OpenVino.
• Improved visual saliency by 21% CNR(Contrast to Noise Ratio) under noisy and extreme lighting conditions for visual surveillance.
• Projects :
• 1. Real Time Video Analytics and Surveillance at the Edge
• 2. Video Management System Back-end

Jun 2018 – Jan 2019 · 7 mos · Gurgaon, India

• Developed & managed 3 AI developers in coding Automatic Document Conversion and Understanding software as a cloud service.
• Key player in design and implementation and responsible for project schedules, milestones, QA and integration C++ micro-services into the overall system acting as a Lead Developer and Scrum Master.
• Introduced fraud check analysis for data verification of soft copies of bank statement.
• Designed scalable & modular OCR Data Extraction, segmentation and entity recognition RESTful services for scanned files for document.
• Coordinated and directed code releases and code reviews.
• Consistently met deadlines and requirements for all production work
• Projects:
• 1. Data Pipeline
• 2. Bank Statement Analyzer
• 3. CoinTurk - Mechanical Turk for semi automatic data annotation and correction

Jan 2016 – Jun 2018 · 2 yrs 5 mos · Gurgaon, India

• Enhancement, Detection & OCR of complex unstructured & structured text from scanned documents.
• Extend FRCNN for page layout detection. Also implemented DeepDeSRT for table detection.
• Image reconstruction and extraction of tables, Automatic Page Layout Detection with 97.9% accuracy.
• Eliminated data entry by using machine learning techniques, boosting the productivity from 13000 ppm to more than 1.5L ppm.
• Devised Tesseract OCR CNN LSTM online model training & evaluation.
• Designed and coded micro-services in C++ & MEAN Stack semi-automated data extraction and annotation service for fast and unbiased correction and labelling of textual images to help train OCR and minimize Concept Drift of ML model.
• Applied semantic search, NER & Sequence Modelling for identifying & capturing details in KYCs.
• Optimized code for performance and reliability, multi-threading at loop levels with data flattening techniques for 3x speedup.
• Projects:
• 1. Financial Documents Digitization and Analysis
• 2. Automatic filling of KYC forms with data extracted from ID cards
• 3. CoinTurk – Data Collection and Annotation for OCR

Feb 2014 – Jan 2016 · 1 yr 11 mos

• VisioLucid(India) is Back Office for VisioIngenii, United Kingdom.
• Managed a cross-functional 6 member team in 2 locations (London and Bangalore), ranging from entry-level to MS developers and closely collaborated with business development, data analysis, operations and marketing teams.
• Assisted in developing Real Time Crowd Detection and Modelling C++ API for automated video analytics using Image Processing, application of FHoG, gradient feature tracking with Kernelized Correlation Filters and Background Modelling.
• Implemented & evaluated OMEGA model for human Head-Shoulder detection, 72% accuracy in cluttered environments.
• Implemented Face Recognition using High dimensional Local Binary Patterns, reduced feature space from 75K to 400 features and verification accuracy of 83.1%, against Labelled Faces in the Wild dataset.
• Projects :
• 1. Crowd Detection and Management through Video Analytics
• 2. Face Recognition with SVM and High Dimensional Local Binary Patterns for London School of Facial Orthotropics
• 3. Real Time Online Video Stabilization
• 4. Automatic Number Plate Recognition in Videos

Oct 2012 – Jan 2014 · 1 yr 3 mos

• Developed state-of the-art Real Time Fire and Smoke Detection for non IR camera sensors with a detection and tracking accuracy of 97%, using spatio-temporal and wavelet analysis.
• Assisted in developing colour invariant marker based augmented reality engine based on object localization using ORB feature detection and Extended Kalman Filter tracking, backend C++ and model rendering using OpenGL.
• Analyzed MRI DICOM images in Frequency Domain to segment tumours.
• Applied for Imagery Library for Intelligent Detection Systems Certification through Home Office- UK.
• Designed Media Layer for Augmented Reality Engine in OpenGL
• Assisted in developing Sketch to Face Recognition and improved the accuracy from 61% to 78% in generating closer matches against the drawn sketches using enhanced local binary patterns.
• Projects:
• 1. Real Time Fire & Smoke Detection onboard for UAV
• 2. Colour Invariant Marker Based Augmented Reality Engine
• 3. Sketch 2 Face Recognition
• 4. Colour Invariant Marker Based Augmented Reality Engine

May 2012 – Oct 2012 · 5 mos · Brighton

• Analysis of various stimuli used in experiments to study body language
• Mathematical representation of data using MS Excel and MATLAB
• Signal Processing using MATLAB to perform transformations and regression analysis
• Implementing various algorithms to optimize results and minimize errors

Sep 2011 – Jan 2012 · 4 mos · University of Sussex, United Kingdom

• Analysed and tabulated the entry requirements for the 1994 group of institutions in UK
• Involved data extraction from various sources
• Timely delivered reports.

Jan 2010 – Apr 2010 · 3 mos · Greater Delhi Area

• As a Field-Engineer supervised the installation and serviced EPBAX systems
• Programmed EPBAX CPU using the OMC software according to the client’s specifications
• Designed the layout for carefully laying out the cables and conducted the final tests
• Inspected the splicing and integration of cables and connectors
• Identified and analyzed the programming faults and repaired PCB cards in the OXO cabinet