Anant Suresh Sakhare

Professional Summary

• Built & deployed real-time event analytics across 2,500+ cameras in 100+ locations, enabling live monitoring and actionable insights, improving peak-hour delivery efficiency by 20% through optimized routing and congestion reduction.
• Automated dataset workflows using synthetic data, auto-annotation (DINO/YOLO + CLIP), and web scraping, reducing preparation time by 35%.
• Contributed to open-source projects through model conversion, tooling, and documentation.

Technical Skills

Languages: Python, SQL

Frameworks: Numpy, Pandas, scikit-learn, TensorFlow, PyTorch, OpenCV, LangChain, FastAPI, Keras, Streamlit, DeepStream, pymongo, Hugging Face, ffmpeg

Developer Tools: Docker, MongoDB, LanceDB, Kubernetes, Git, Linux Shell, Triton Server, ml-flow, Code Profilers

Certifications: IBM Data Science Professional, Python for Everybody

Professional Experience

Data Scientist - Computer Vision
Apptware Solution LLP  |  Pune, MH, India  |  Nov 2025 - Current

• Working on builing a intelligent AI system for crowd management. ( on going)

Computer Vision Engineer I & II
Wobot Intelligence Private Limited  |  New Delhi, India  |  Jul 2022 – Nov 2025

• Designed and deployed a multi-camera vehicle and person tracking system across 2500+ live cameras in 100+ locations, enabling real-time traffic and crowd monitoring. Improved peak-hour delivery efficiency by 20% through faster routing and reduced congestion.
• Trained and optimized classification and detection models (ResNet, Inception, EfficientNet, EfficientDet, YOLO) with advanced retraining, augmentation, and negative mining, reducing false detections from 15% to 0.1%, cutting manual cross-verification workload by 85%.
• Trained and fine-tuned Re-ID models with ResNet50 and enhanced CLIP variants using prompt-learning methods (LoOp, CoOp, CoCoOp, MAPLE), boosting cross-camera identity tracking accuracy for surveillance, delivery optimization, and crowd analytics.
• Built CLIP-based pipelines with LanceDB for semantic image–text retrieval, and fine-tuned Gemini on Vertex AI for event grading of CNN outputs, reducing false IDs by 25%.
• Led mentorship for interns on CV projects and deployment workflows; contributed to hiring interviews.

Data Engineer I
Wobot Intelligence Private Limited  |  New Delhi, India  |  Dec 2021 – Jun 2022

• Developed an auto-annotation pipeline using DINO and YOLOv8 for object detection and CLIP for classification, reducing dataset preparation time by 35% and accelerating model training.
• Created synthetic datasets for computer vision models, reducing labeled data needs by 75% and speeding delivery.

Data Scraping Intern
Wobot Intelligence Private Limited  |  New Delhi, India  |  Aug 2021 – Dec 2021

• Used unsupervised ML algorithms to accelerate data collection, reducing processing time by 50%.
• Ensured quality and consistency in annotated datasets for CCTV-based use cases.
• Scraped high-quality image data from web and YouTube sources for training and validation.

Summer Intern — Industrial Engineering Lab
Indian Institute of Technology (BHU), Varanasi  |  Varanasi, India  |  May 2018 – Jun 2018

• Designed and implemented a Vehicle Routing Problem (VRP) model, reducing transportation costs by 7% and delivery time by 20 minutes for a dairy manufacturer.
• Applied k-means clustering & Cheapest Link Algorithm to optimize routes and resource use.

Open Source Contributions

• Enhanced Trolo project with TensorRT and ONNX conversion support, adding pre/post-processing scripts.
  Repository | PR Engine file format | PR ONNX format
• Developed detailed documentation for Atlas AI project using Docusaurus for easy onboarding.
  Documentation PR

Education

Bachelor of Engineering in Mechanical Engineering
Brahmdevdada Mane Institute of Technology (BMIT), Solapur University, Solapur, MH, India
Aug 2016 – Jun 2019

Leadership & Achievements

• President, MESA, BMIT, Solapur: Hosted 7 interdepartmental technical and artistic events with 2,000+ participants and a team of 7. Established a student help desk for queries.
• Lead Coordinator, AbhiVyakti-2k19, BMIT, Solapur: Managed a national-level technical and artistic competition with 28 events, 2,000+ students, and team of 7 committee members.
• Competitor, zonal athletics: 4 × 100 meters relay and 200 meters running competition.