MOHAMMED MUZAMMIL PESHIMAAM
Sr. Conversational AI Developer | M.S. AI & ML @Woolf University, Canterbury, UK | Winner Smart India Hackathon 2020 | Patent Holder.
Hyderabad
4+ years of work experience
Dec 20256 mosMay 2026
On-site
- Developed and deployed a new prompt techniques that significantly reduced token usage. This optimization directly reduced operational costs while maintaining high-quality, accurate, and fast LLM outputs for key business systems.
- Engineered end-to-end OCR beautifier converting Amazon OCR JSON outputs to pixel-perfect image reconstructions using mathematical algorithms and geometric transformations — Increased performance of the model by 20%.
- Built end to end pipeline handling millions of records from postgres SQL via Celery, symphony, async/await patterns, and queue-based processing to generate business statistics — reducing latency by 70%, enabling real-time business intelligence
Oct 20241 yr 2 mosNov 2025
Hyderabad
- Developed and deployed chatbots and voicebots using Generative Playbooks, Flows in Dialogflow CX/ Conversational Agents, and designed architectures using LucidCharts. optimizing user journeys.
- Developed custom DFCX Tools using Cloud Functions that interfaced with External APIs to efficiently retrieve and supply necessary data parameters directly into Dialogflow CX, enriching conversational capabilities.
- Built an AI-powered LLM solution (Agentic AI, LangGraph, Gemini 2.5 flash) to automate instruction and example generation for playbooks, achieving a 40% increase in creation efficiency and reducing average development time from 2 weeks to 1 week.
- Created and implemented an innovative prompting technique,(Condition + Action), to effectively reduce LLM hallucinations and overfitting with training examples, substantially improving model reliability by 70%.
- Developed and managed a structured datastore of diverse URLs, integrating it via a Retrieval Augmented Generation (RAG) approach to generate more accurate and contextually relevant bot responses, thereby minimizing manual refinement efforts.
- Proactively engineered comprehensive test strategies beyond standard UAT, utilizing Dialogflow CX's diagnostic tools and anticipating integration points with the broader Google CCAI suite (e.g., Agent Assist, Insights) to ensure end-to-end solution robustness and optimize for downstream analytics
Dec 20222 yrs 1 moDec 2024
- Developed and deployed chatbots using Flows in Dialogflow CX and designed architectures using LucidCharts
- Analyzed design deficiencies and engineered robust internal and common fallback strategies for out-of-scope user utterances, ensuring a more intuitive and seamless user interaction.
- Analyzed Dialogflow CX logs and user interaction data to identify and address issues related to API failures, model inference, prompt injection and refined using NLU’s Intents, entities, prompting strategies, training examples, event handlers, generators, adding common guardrails and cyber guardrails.
- Prioritized Responsible AI by integrating comprehensive guardrails and conducting rigorous testing to ensure empathetic, respectful, and trustworthy bot responses, enhancing the user experience.
- Actively supported User Acceptance Testing (UAT) by creating initial test cases, diligently fixing bugs identified during testing, and providing technical guidance on complex conversational flows.
- Served as a key developer during hypercare and post-launch phases, providing rapid bug-fix resolution in production to maintain solution stability, performance, and alignment with business objectives.
- Articulated strategic design changes and project outcomes to international stakeholders, fostering cross-functional alignment for successful chatbot deployments.
- Independently managed code merging processes and facilitated the seamless movement of code and configurations between different Dialogflow CX agents, ensuring development agility and consistency.
Feb 20243 mosApr 2024
Fine – tuned the model using RLHF, leveraging human feedback to refine large language models. Crafted and optimized prompts to guide models towards relevant answers, contributing to the advancement of AI algorithms.
Developed and iteratively refined over 100+ prompts based on user feedback, enhancing model performance and accuracy.
Collaborated with peers to share insights and best practices in prompt design and optimization. Contributed to the advancement of prompting methodologies by integrating cutting-edge techniques such as Few Shot Prompting, RAG, COT, and Self Consistency into task workflows.
Promoted to reviewer within a mere 2-week timeframe, demonstrating rapid progression and proficiency.
Oct 20223 mosDec 2022
Remote
- Developed an interactive chatbot leveraging AI semantic search and Created a personalized resume maker for students using Generative AI (OpenAI), tailoring solutions to meet unique individual requirements.
- Instructed educational sessions for over 60+ students on Deep Learning Projects like OpenCV, RNN, CNN, LSTM, Attention etc.., providing practical insights and facilitating a deeper understanding of the subject.
- Produced datasets for student projects, resulting in significant recognition, with top 3 projects garnering 13k, 7k, 4k views and 2k, 1k, 800 downloads respectively, each earning a bronze medal.
Nov 202111 mosSep 2022
Remote
- Trained and mentored over 100+ students in Data Structures Algorithms
- Developed and delivered an engaging and interactive DSA curriculum that was aligned with industry standards and best practices.
- Solved over 600+ problems on Leetcode, 200+ problems on GFG.
Education
Master of Science - MS, Artificial Intelligence & Machine Learning
Woolf
20222 yrs2024
Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering
CMR College of Engineering & Technology
20184 yrs2022
Join MOHAMMED and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss