Sr Sde - Data Science
Current
Sep 2024 - Present
AeroLeads people directory · profile
Ajith Balakrishnan Email & Phone Number
Multi-Model LLM | Gen AI | Computer Vision | NLP | TensorRT | Embedded Edge |
at
Moder
LinkedIn matched
✓ Verified July 2026
3 data sources
Profile completeness 86%
Current company
Role
Multi-Model LLM | Gen AI | Computer Vision | NLP | TensorRT | Embedded Edge |
Location
Kerala, India
Who is Ajith Balakrishnan? Overview
A concise factual answer block for searchers comparing this professional profile.
Quick answer
Ajith Balakrishnan is listed as Multi-Model LLM | Gen AI | Computer Vision | NLP | TensorRT | Embedded Edge | at Moder, based in Kerala, India. AeroLeads shows a matched LinkedIn profile for Ajith Balakrishnan.
Ajith Balakrishnan previously worked as Sr SDE - Data Science at Moder and Associate Research staff member at Slk. Ajith Balakrishnan holds Mtech, Robotics And Automation from College Of Engineering, Trivandrum.
Company email context
Email format at Moder
This section adds company-level context without repeating Ajith Balakrishnan's masked contact details.
Moder
Review company-level records connected to Ajith Balakrishnan before choosing the right outreach path.
Profile bio
About Ajith Balakrishnan
I am a seasoned computer vision engineer deeply passionate about advanced technologies, including computer vision (CV), natural language processing (NLP), edge devices, and Generative AI (GenAI). I have extensive experience with CV frameworks like TensorFlow and PyTorch and edge devices such as Jetson, DeepStream, OAK, and TensorRT. My work with large language models and Multi-Model Gen AI solutions fuels my drive to innovate in the tech industry.My master's thesis centered on 3D Model Reconstruction from 2D Images. Despite the challenges, our innovative CNN architecture and training algorithm achieved state-of-the-art results. The solution integrated a visual auto-encoder with self-attention, followed by a 3D refiner network, and used a unique training algorithm called Joint Train Separate Optimization. This project began by addressing the difficulties of matching image features across 2D views using computer vision. I am currently working on publishing these research findings.Professionally, I have specialized in Document Information Extraction from various scanned documents, both structured and unstructured. I developed OCR frameworks for both handwritten and machine-printed documents, leading the creation of multi-modal transformers (VLM)-based OCR for robust extraction across diverse formats. My expertise extends to deep learning solutions using edge devices like Jetson Nano and Jetson Xavier.At V-Labs, I focused on object detection using Jetson devices, conducting quantization, pruning, and QAT experiments with YOLO and SSD models. I have hands-on experience with advanced language models such as GPT-4, Mistral, and LLAMA for Retrieval-Augmented Generation (RAG) and Sentiment Analysis projects. I have fine-tuned these models using the HuggingFace library and optimization algorithms like PEFT-LORA, SFT, and DPO, enhancing my ability to build and implement practical AI solutions. These experiences have honed my skills in developing reliable algorithms and ensuring the accuracy, efficiency, and reliability of AI systems.My research interests lie in Multi-Modality LLM, Edge AI, CV, and Reinforcement Learning (RL). The rise of LLMs and multi-modality techniques presents a promising frontier for innovative solutions to complex real-world problems. Recent research highlights potential gains in speed, accuracy, transparency, and explainability in the learning process. These advancements could significantly improve the deployment and usability of AI systems in everyday applications, a field I am eager to explore further in my career.
Current workplace
Ajith Balakrishnan's current company
Company context helps verify the profile and gives searchers a useful next step.
9 roles
Ajith Balakrishnan work experience
A career timeline built from the work history available for this profile.
Associate Research Staff Member
Apr 2024 - Aug 2024
Associate Research Staff Member
Project: Lockbox- Automated Information Retrieval from Scanned DocLed the development and deployment of an Information extraction solutionbased on Python with a six-sigma confidence level (3.4 errors per million).Integrated ViT encoder and BART decoder as a pipeline architecture toperform Information Extraction for handwritten documents.Proposed FasterRCNN-based Chargrid algorithm to improve localization F1-score by 13.45 on documents.Developed an advanced OCR… Show more Project: Lockbox- Automated Information Retrieval from Scanned DocLed the development and deployment of an Information extraction solutionbased on Python with a six-sigma confidence level (3.4 errors per million).Integrated ViT encoder and BART decoder as a pipeline architecture toperform Information Extraction for handwritten documents.Proposed FasterRCNN-based Chargrid algorithm to improve localization F1-score by 13.45 on documents.Developed an advanced OCR algorithm based on a vision-languagetransformer, resulting in a 17% reduction in CER compared to CRNNalternatives.Project: SentixAi- An Intelligent Finance Advisory PlatformDeveloped an LLM fine-tuning framework using PEFT-LoRA and QLoRA, withsupporting trainers like SFTT, DPO, and ORPO methods for ABSA & IR tasks.Developed RAG pipeline using Python & LangChain with GPT4 and variousLlama models for intelligent document extraction.Studied Instruction fine-tuning and soft prompt-tuning on compound ABSAtasks using LLMs FlanT5, Llama2, and Mistral.Project: Face Recognition: A Distributed Edge Computation FrameworkDeveloped a Multi-stream object detection pipeline using Python,DeepStream & Jetson EdgeAI with DeepSORT and IOU trackers.Finetuned YOLO-v4 and SSD models with QAT using NVIDIA-TAO andachieved an accuracy of 0.97+ for Near-field and Far-field detection datasets(Custom)Developed ArcFace Face Recognition model (Accuracy - 0.965) on LFW, CFP-FP & AgeDB datasets.Compared the performance of CNN with transformer-based object-detectionmodels like DETR and SwinTransformer Show less
Jun 2022 - Apr 2024
Mar 2021 - Jun 2022
Project Engineer
Sep 2019 - Apr 2021
Feb 2019 - Jul 2019
Intern Robotics
Myrobo Robotics Academy
Ros, Deep Learning, Embedded System.
May 2018 - Feb 2019
Apr 2017 - Jul 2017
Embedded Engineer
Arvin Technologies
Embedded Systems, Edge devices, Computer vision
Oct 2015 - Apr 2017
2 education records
Ajith Balakrishnan education
Mtech, Robotics And Automation
B Tech, Electronics And Communication
FAQ
Frequently asked questions about Ajith Balakrishnan
Quick answers generated from the profile data available on this page.
What company does Ajith Balakrishnan work for?
Ajith Balakrishnan works for Moder.
What is Ajith Balakrishnan's role at Moder?
Ajith Balakrishnan is listed as Multi-Model LLM | Gen AI | Computer Vision | NLP | TensorRT | Embedded Edge | at Moder.
Where is Ajith Balakrishnan based?
Ajith Balakrishnan is based in Kerala, India while working with Moder.
What companies has Ajith Balakrishnan worked for?
Ajith Balakrishnan has worked for Moder, Slk, V-Labs, Wipro Limited, and Myrobo Robotics Academy.
How can I contact Ajith Balakrishnan?
You can use AeroLeads to view verified contact signals for Ajith Balakrishnan at Moder, including work email, phone, and LinkedIn data when available.
What schools did Ajith Balakrishnan attend?
Ajith Balakrishnan holds Mtech, Robotics And Automation from College Of Engineering, Trivandrum.
Find 750M verified contacts
Search by job title, company, industry, location, and seniority. Export verified B2B contact data when you need it.
Start free trialPeople with similar names
Check these profiles if this is not the Ajith Balakrishnan you were looking for.
View similar profiles
