About Me
AI/ML Engineer building agentic products, RAG systems, and computer vision that ships.
I work where AI stops being a demo and starts becoming product: backend systems, model pipelines, semantic search, agent orchestration, and the last-mile details that make users trust the thing.
Right now my focus is Neighborly, a social platform for real-world neighbor discovery, local activities, events, and community engagement.
Before that, I shipped production AI across computer vision, video automation, SEO intelligence, semantic search, recommendations, and synthetic data. The common thread is simple: fast prototypes, clean systems, and AI that earns its place in the workflow.
Community discovery, activity matching, events, and local engagement.
LangGraph, RAG, edge CV, synthetic data, and semantic retrieval.
Prototype quickly, harden the useful parts, and keep the product moving.
Facts
A few human details behind the builder mode. I like making things on screens, but I also like getting away from them.
I paint when ideas need color, texture, and a break from terminals.
I backpack light, chase mountain air, and collect stories from the road.
I travel for long walks, local food, quiet views, and unexpected detours.
I build
apps, AI systems, communities, and tiny experiments that turn into bigger ones.
Skills
The stack I reach for when the job moves from idea to shipped system.
Deepankar Sharma
AI/ML Engineer | Generative AI | Agentic Systems | Computer Vision
- deepankarsharma2003@gmail.com
- linkedin.com/in/ideepankarsharma2003
- github.com/ideepankarsharma2003
- Nainital, Uttarakhand, India
Resume
Latest snapshot: AI/ML engineer, generative AI builder, and product-minded systems developer. This section is formatted to print cleanly, no separate PDF needed.
Summary
Deepankar Sharma
AI/ML engineer building production generative AI, agentic automation, RAG, semantic search, and computer vision systems. Experienced across early-stage product ownership, backend engineering, model optimization, data pipelines, and cloud deployment.
- Nainital, Uttarakhand, India
- deepankarsharma2003@gmail.com
- linkedin.com/in/ideepankarsharma2003
Experience
Computer Vision Engineer
Nov 2024 - 2026
i2V Systems, Gurugram
- Optimized person re-identification and attribute models for production inference with ONNX, OpenVINO, and edge deployment.
- Led video synopsis development and automated training pipelines, reducing the label-to-train cycle by 65%.
- Built synthetic data workflows using ControlNet, inpainting, and generative image composition.
ML Developer / Senior ML Developer
Aug 2023 - Oct 2024
Warewe Consultancy, Gurugram
- Scaled Shotify.ai on Django/GCP to support thousands of daily AI video generations for short-form social content.
- Designed Neo4j-backed RAG and semantic-search systems using GTE/BGE embeddings for production knowledge retrieval.
- Fine-tuned LLaMA-2/3 with LoRA and shipped GenAI systems with Stable Diffusion, SAM, Whisper, LangChain, and SEO automation.
Education
Master of Computer Applications (MCA)
2023 - 2025
Graphic Era University (ODL)
Focus: Machine Learning, Data Science, and applied AI systems.
Bachelor of Computer Applications (BCA)
2020 - 2023
Graphic Era University, Haldwani
Gold Medalist | SGPA: 9.4/10
Selected Projects
Neighborly
Current Build
Social platform for real-world neighbor discovery, activity matching, event participation, and community engagement.
RedSerp, VitalPeptides & Open Source ML
Product Systems
- RedSerp: Reddit trend intelligence with automated PostgreSQL ingestion, topic discovery, and ML-based content optimization.
- VitalPeptides: Personalized recommendation engine using regression analysis and customer signals.
- Published 17+ models and datasets on Hugging Face with 20k+ downloads across applied ML experiments.
Active AI Builds
Founding AI Engineer
Nov 2025 - Present
Redegades
- Architecting an AI Co-CEO decision-intelligence platform using LLMs, LangGraph, RAG, Fireflies, and multi-agent orchestration.
- Built OpenClaw and Hermes based agentic automation combining semantic search, tool use, and cross-application actions.
AI & Lead Developer
Nov 2025 - Present
LabThunder
- Leading a multi-agent lab operations platform with LangGraph, RAG, ML analytics, and workflow optimization.
- Designed AI workflows for equipment discovery, task routing, decision support, and operational analytics.
Certifications
Selected Certificates
- Google Data Analytics Professional Certificate - Google
- Machine Learning Specialization - DeepLearning.AI
- Deep Learning Specialization - DeepLearning.AI
- Machine Learning with TensorFlow on Google Cloud - Google Cloud
Portfolio
Flagship products, production systems, older experiments, and the proof-of-work trail behind them. The front row is what I would show first in a room full of builders.
- Flagship
- Projects
- Kudos
- Certifications
- Notes
Current Build
Business Copilot
Redegades
AI business copilot for decision intelligence, meeting memory, and cross-application automation.
Details
Multi-Agent Ops
AI Video
Shotify.ai
Short-form video generation and publishing pipeline built on Django and cloud infrastructure.
Visit
Trend Intel
RedSerp.com
Reddit trend intelligence with automated ingestion, topic discovery, and ML content insights.
Visit
Content Agent
Paperbot.AI
GPT-4, Whisper, LangChain, and CrewAI system for turning voice and notes into long-form content.
VisitServices
I build practical AI fast: agents, retrieval systems, vision models, backend APIs, and the ops needed to keep them useful.
Agentic Workflows
LangGraph, tool use, workflow routing, memory, and automation that survives messy tasks.
RAG & Semantic Search
Grounded retrieval, embeddings, graph-backed search, and answers that can be trusted.
Recommendation Engines
Customer signals, regression, product matching, and ranking logic tuned for outcomes.
Computer Vision
Re-ID, attributes, video synopsis, synthetic data, and edge deployment for real-time systems.
Testimonials
Don’t just take my word for it—here’s what clients, collaborators, and teammates have said about working with me.
Contact
Have a serious AI problem, a weird prototype, or a product that needs to get real? Let’s talk.
I am especially interested in agentic systems, computer vision, community tech, and useful automation.
Location:
Nainital, Uttarakhand, India